Professional Documents
Culture Documents
Sg244656 VTAM Subarea To APPN
Sg244656 VTAM Subarea To APPN
http://www.redbooks.ibm.com
This book was printed at 240 dpi (dots per inch). The final production redbook with the RED cover will
be printed at 1200 dpi and will provide superior graphics resolution. Please see “How to Get ITSO
Redbooks” at the back of this book for ordering instructions.
SG24-4656-01
IBML
SG24-4656-01
International Technical Support Organization
May 1998
Take Note!
Before using this information and the product it supports, be sure to read the general information in
Appendix L, “Special Notices” on page 263.
When you send information to IBM, you grant IBM a non-exclusive right to use or distribute the information in any
way it believes appropriate without incurring any obligation to you.
Copyright International Business Machines Corporation 1996 1998. All rights reserved.
Note to U.S. Government Users — Documentation related to restricted rights — Use, duplication or disclosure is
subject to restrictions set forth in GSA ADP Schedule Contract with IBM Corp.
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
The Team That Wrote This Redbook . . . . . . . . . . . . . . . . . . . . . . . . . vii
Comments Welcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Scope and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 APPN Terminology and Functionality . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Types of SNA Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.2 Types of SNA Node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.3 Some Confusing Terminology . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.4 Types of Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.5 Network Control Sessions . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.6 Dependent LU Support in APPN . . . . . . . . . . . . . . . . . . . . . . 9
1.2.7 HPR Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2.8 Types of SNA Data Units . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2.9 APPN Directory Services . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.10 APPN Topology and Class of Service . . . . . . . . . . . . . . . . . . 13
1.2.11 Other APPN Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.2.12 APPN Session Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Contents v
Appendix E. VTAM APPN Migration Definitions . . . . . . . . . . . . . . . . . 237
E.1.1 Transmission Group Profile Default Table IBMTGPS . . . . . . . . . 237
E.1.2 APPN Entries in ISTINCLM IBM Default Logmode Table . . . . . . . 239
E.1.3 RAAMDLS Customized MODEL Major Node . . . . . . . . . . . . . . 241
Appendix G. VTAM Definitions with RAA as ICN and RAS as MDH . . . . . 245
G.1 VTAM RAA Definitions Updated for MDH . . . . . . . . . . . . . . . . . . 246
G.1.1 Configuration List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
G.1.2 CDRM Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
G.1.3 ANNC Definitions from RAA to RAS . . . . . . . . . . . . . . . . . . . 247
G.2 VTAM RAS Definitions Updated for MDH . . . . . . . . . . . . . . . . . . 248
G.2.1 Start Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
G.2.2 Configuration List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
G.2.3 CDRM Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
G.2.4 ANNC Definitions from RAS to RAA . . . . . . . . . . . . . . . . . . . 250
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
This redbook is the first of two volumes containing detailed coverage of the
migration of a subarea network to an APPN network. It focuses on the migration
steps and requirements that can be used as guidelines to accomplish the
migration in a simple and constructive manner. The first volume covers the
implementation of the basic VTAM APPN functions including border node, VR-TG
and network management. The second volume (SG24-5204) completes the
coverage of a typical customer′s network using HPR, DLUR and APPN/HPR
routers such as 3746, 2216 and 2210. In each case tested examples and
definitions illustrate the theory.
The first eight chapters in this volume are based on the first edition of this book
(SG24-4656-00). They have been rewritten to improve the explanations and
clarify the examples. They have also been updated with information on the
latest release of eNetwork Communications Server for OS/390.
Some knowledge of SNA subarea networks and familiarity with the functions,
terms and data flows of APPN networks is assumed.
Jerzy Buczak
Systems Management and Networking ITSO Center, Raleigh
Johnathan Harter
CS OS/390 Development, Research Triangle Park
John Parker
IBM United Kingdom, Manchester
The original version of this book was the result of a project designed and
managed by:
Michael Li
Systems Management and Networking ITSO Center, Raleigh
Anar Kermali
IBM Canada
Rabinder Mokha
IBM Australia
Ken Oag
National Australia Bank, Australia
Thanks to the following people for the invaluable advice and guidance provided
in the production of both editions of this document:
Pepe Boo
IBM Spain, Madrid
Roy Brabson
CS OS/390 Development, Research Triangle Park
Fran Collins
Networking Software Sales and Opportunity Solutions, Dallas
Gwendolyn Dente
Washington System Center, Gaithersburg
Jim Fletcher
CS OS/390 Design, Research Triangle Park
Nancy Gates
Washington System Center, Gaithersburg
Johnathan Harter
CS OS/390 Development, Research Triangle Park
John Klonowski
CS OS/390 Development, Research Triangle Park
John Parker
IBM United Kingdom, Manchester
Sam Reynolds
CS OS/390 Development, Research Triangle Park
Juan Rodriguez
Systems Management and Networking ITSO Center, Raleigh
Carla Sadtler
Systems Management and Networking ITSO Center, Raleigh
Suvas Shah
Communications Manager and Communications Server Architecture,
Research Triangle Park
Ann Wright
CS OS/390 Development, Research Triangle Park
Comments Welcome
Your comments are important to us!
Preface ix
x Subarea to APPN Migration: VTAM and APPN Implementation
Chapter 1. Introduction
The history of APPN within VTAM is relatively new. In 1993, IBM introduced
APPN into the mainframe environment with VTAM V4R1 and NCP V6R2. There
were many restrictions, which meant that some customers did not feel the need
to migrate to or implement VTAM V4R1. They could not see any real benefits.
For example, channel-to-channel connections were not supported. Dependent
LUs had to be adjacent to the subarea network. SNI connections had to be used
at all network boundaries. Route calculation was haphazard between the APPN
network and the subarea network.
VTAM V4R2 addressed most of these issues in terms of APPN. Border node
functions, connection networks, APPN host-to-host connection, VR-TG, and
Dependent LU Requester/Server functions were all introduced with this release.
The latest generally available release at the time of writing is VTAM V4R4.1,
which is actually part of eNetwork Communications Server for OS/390 Release 5
and not available separately. CS OS/390 Release 5 provides full support for
high-performance routing (HPR), which was introduced for some configurations
by V4R3 and developed fully by V4R4. HPR utilizes a rapid transport protocol
(RTP) connection to transport session traffic between session endpoints. HPR
routes can traverse existing subarea routes using VR-based TGs between
intermediate nodes.
There are many different products and platforms which support APPN and can
therefore be integrated into an APPN network. These include platforms made by
many companies other than IBM. The list includes offerings from companies
such as Apple Computers, Brixton Systems, Cabletron Systems, Cisco Systems,
Bay, 3Com, Nortel, Unisys and many others. The list of non-IBM suppliers for
LEN platforms, which can participate in APPN networks, is even longer. The
user should determine the level of support offered by these vendors.
The main objective of this book is to outline the most important steps required to
migrate an existing subarea network to a mixed subarea and APPN network. It
is envisaged that mixed subarea/APPN/HPR environments will exist for some
years to come. These three SNA protocols can coexist and be intermixed quite
The objective of this book, therefore, is to define a subarea network and migrate
it using a number of easy-to-follow steps to a mixed subarea and APPN network.
These steps will help readers to migrate their own networks in a simple and
constructive way. Although the network which we are working with will not
match every reader′s network, whether due to size or actual devices used, we
simulate as many scenarios as possible. The migration will include the use of
APPN functions such as border node capability, VR-TG, ANNC, and SSCP
Takeover/Giveback. A companion book (SG24-5204) will cover extensions to the
base VTAM APPN network including HPR, DLUR/S and APPN routers such as
3746, 2216 and 2210.
Note
This pair of redbooks covers migration from subarea to APPN, and then to
HPR, in two separate volumes. We must emphasize that this is purely to
help the reader understand APPN and HPR, and is not intended as a
recommendation to perform the migration in two phases. The difference
between APPN and HPR is small compared to the difference between
subarea and APPN. You may find it advantageous to let VTAM′s defaults
take you all the way to HPR as you implement APPN.
A subarea node can only support APPN connections as LEN connections. LEN
support is the same as provided in V3R2 to V3R4.2 of VTAM.
Each resource in a subarea network is uniquely identified by its NetID and its
resource name.
Chapter 1. Introduction 3
An APPN network may be divided into subnetworks, although the APPN scheme
is more flexible than the subarea scheme. APPN resources are also identified
uniquely by their NetID and resource name.
LEN nodes can directly communicate with other LEN nodes without requiring the
services of an SSCP. An SSCP is required if LEN nodes are to communicate
with each other across a subarea network, but the communication with the SSCP
is done by the subarea boundary function and is transparent to the LEN nodes.
The great disadvantage of LEN connections, apparent in both subarea and APPN
networks, is that no network control functions (such as searching) can take place
over them. As a result, much more predefinition is required than for the
equivalent APPN or subarea connection.
When HPR is in use, LFSIDs are not used for the HPR portion of the sessions.
Instead, ANR labels are used by each node to route HPR network layer packets
(NLPs).
A network node supports CP-CP sessions, but need not support either
SSCP-SSCP sessions or subarea connections. VTAM, when defined as a “pure”
network node, is not able to own or activate NCPs and has no subarea number
defined. When VTAM is defined as both a subarea node and an APPN network
node, it can own and activate NCPs. Such a VTAM, in conjunction with one or
more NCPs, is known as a composite network node (CNN) and appears as a
single APPN node.
Chapter 1. Introduction 5
connected to other APPN nodes, LEN nodes, and subarea nodes. An MDH
cannot perform intermediate routing, except wholly within a subarea network,
although its locally attached LEN nodes and dependent LUs can establish
sessions with resources in other APPN or subarea domains.
Two types of border node are defined in the APPN architecture: peripheral
border node and extended border node.
1.2.3.1 Domain
The concept of a domain is different in subarea and APPN networks. In a
subarea network, it means a VTAM and all the resources it owns (with which it
has SSCP-PU or SSCP-LU sessions). In an APPN network, it means a network
node and all the resources it serves; this therefore includes resources on end
nodes served by the network node.
Chapter 1. Introduction 7
1.2.4.2 Virtual Route-Based Transmission Group
A virtual route-based transmission group is a VTAM-specific routing function that
enables two APPN-capable VTAM nodes to define an APPN connection between
them across a subarea route . Such an APPN connection is reported to the APPN
topology database in the usual ways, but has a unique TG number of 255.
One advantage of VR-TG is that it provides full APPN function across an existing
subarea network, thus alleviating the need to redefine connections when a
speedy migration is desired. A VR-TG allows APPN LU-LU sessions to be set up
using subarea virtual routes across the subarea portion of an SNA network.
Please refer to Chapter 7, “VR-Based TG Implementation” on page 131 for a full
description.
Within a subnetwork, in order to establish an LU-LU session the SSCPs that own
the LUs must be in session with each other. If the LUs are in different
subnetworks the SSCPs that own them must be able to communicate with each
other using a chain of SSCP-SSCP sessions. The links in the chain are the
gateway SSCPs that control the subnetwork boundaries.
For dependent LUs, the SSCPs must also establish SSCP-PU and SSCP-LU
sessions before an LU-LU session can be set up. Independent LUs are those
which do not require SSCP-LU sessions. This leads to the major difference
between a type 2 and a type 2.1 node; the latter can support independent LUs as
well as dependent while the former can only handle dependent LUs.
You are not required to have CP-CP sessions between each and every node in
the network. Direct CP-CP sessions are not required between the endpoints of
an LU-LU session. A network node can have CP-CP sessions with any other
network node to which it has an APPN connection. It also maintains CP-CP
sessions with each of the end nodes it serves. End nodes can be attached to
several network nodes but can only establish CP-CP sessions with one network
node server at a time. CP-CP sessions between end nodes and their network
node servers are used for resource registration and locating destination LUs
when the end node wishes to set up a session. CP-CP sessions between
network nodes are used to register resources with a central directory server,
perform searches for resources, and exchange topology information. Direct
CP-CP sessions between end nodes are never permitted.
A LEN node does not support CP-CP sessions. Therefore, its independent LUs
cannot request an APPN network search, nor can they be found dynamically.
Similarly, because those independent LUs have no SSCP-LU sessions they suffer
the same restrictions in a subarea network. Some predefinition is always
required if an LU on a LEN node is to establish a session.
Chapter 1. Introduction 9
1.2.6.2 Dependent LU Requester
The dependent LU requester function allows an end node or network node to
provide a remote boundary function for dependent LUs. This function relieves
the restriction that dependent LUs and their type 2.0 PUs be directly attached to
the VTAM or NCP boundary function. The DLUR function may reside in the same
node as the dependent LU or be provided by a node adjacent to and upstream
from the dependent LU.
The major advantage of using DLUR and DLUS is that dependent LU sessions
can fully utilize the APPN network. They do not need to flow via a subarea node
adjacent to the dependent LU; they can take the optimum path through the
network as calculated by an APPN network node. They can also take advantage
of HPR.
Any APPN configuration of VTAM V4R4 or above can support RTP, whether
across native APPN connections or VR-TG connections. NCP cannot support
RTP, and the HPR portion of a session cannot terminate in an NCP.
When an RTP connection fails, the RTP endpoint will attempt to find a new route
until the path switch timer expires. The path switch timer duration is modifiable
and can take a different value for each transmission priority.
1.2.8.1 FID2
FID2 is the type of format identifier (FID) in a transmission header (TH) used
between a type 4 or 5 subarea node and an adjacent peripheral (type 2.0 or 2.1)
node. It is also used between adjacent APPN or LEN nodes. The format of a
FID2 header differs depending on whether a type 2.1 node is one of the partners
or not. FID2 refers to the hex digit 2 at the start of the header. A FID2
connection in this book refers to an APPN or LEN connection between two nodes
over which the FID2 PIUs flow.
1.2.8.2 FID4
FID4 is the type of format identifier in the transmission header (TH) used to route
data between subarea nodes. FID4 refers to the starting hex digit of 4 in such a
header. A FID4 connection refers to a subarea connection between two nodes
over which the FID4 PIUs flow.
1.2.8.3 FID5
FID5 is the type of format identifier in the transmission header (TH) used for
sessions flowing over RTP connections and was introduced to accommodate the
new enhanced session addressing algorithm. The FID5 TH contains the 8-byte
session address which is assigned at each RTP endpoint. This session address
is unique per session and direction (PLU-to-SLU and SLU-to-PLU). Each side
assigns the session address that has to be used by the session partner, when
sending data, to address the session. The FID5 PIU is imbedded in the network
layer packet (NLP).
Chapter 1. Introduction 11
1.2.9 APPN Directory Services
A CDS provides a focal point for locating resources in the network. The
resource location information is forwarded to the CDS by the network node
serving the resources, at the request of the actual owner (the NN or a served
EN). The CDS directory is exactly the same as any NN′s directory, except that it
is usually larger.
If an NN searching for a target resource cannot find it in its own directory, the
search request is forwarded to the CDS. If the CDS is unaware of the resource,
the CDS then sends the search to other central directory servers, if they exist in
the network. If the search receives a negative response (resource is unknown)
from these other central directory servers, then a broadcast search is issued by
the original central directory server. If the resource is located by the search, its
location will be added to the CDS database and also to the database of the NN
initiating the search.
If a central directory server is not used, a broadcast search is initiated from the
network node server on behalf of directly attached LUs, LEN-attached LUs or LUs
attached to served end nodes. The search information retrieved is not
communicated to other network nodes. It is stored in the network node server′ s
directory services database. If another network node receives a request for the
same resource and does not have the location information, it will issue another
broadcast search.
You can see from this that a central directory server can bring many benefits to
an APPN network by reducing the numbers of searches. Bear in mind that it is
possible to have more than one central directory server in an APPN network.
In VTAM it is possible to checkpoint and save the TDB to increase the speed of
network restarts. Checkpoints can be issued via a VTAM MODIFY command and
automatically occur when bringing down VTAM with a Z NET or a Z NET,QUICK.
After VTAM is restarted only new information is exchanged. For VTAM to restart
using the checkpointed TDB the start option INITDB should be set to ALL or
TOPO. The default is ALL.
The APPN topology and routing services component consists of the following
three parts:
• Topology database manager
Chapter 1. Introduction 13
• Route selection services
• Class of service manager
In APPN, transmission groups are defined between APPN and/or LEN nodes.
FID4 connections (subarea-subarea) are not included in the APPN topology
database (unless VR-TG is used), while FID2 TG connections are. The
transmission groups are described by APPN characteristics. Some of these are
user-defined, and others are dynamically defined.
Where all the actual TG values fall within the ranges specified by an entry in the
APPN COS table, the weight associated in that entry is the TG weight attributed
to that TG. These TG weights are then aggregated over all possible TGs to form
the route weight. They are used in conjunction with a user-defined node
weighting for each intermediate node on the route. In VTAM this is defined as
the route additional resistance (a value between 0 and 255). The path chosen for
a session is the route with the lowest total (TGs plus nodes) weight.
VTAM V4R1 introduced the new VTAM definition statement TGP. Used in
conjunction with the PU parameter TGP= it allows TG characteristics to be
easily and conveniently defined for each connection. There are also new VTAM
commands for displaying and modifying transmission group profile
characteristics. See VTAM Resource Definition Reference, SC31-8565 and VTAM
Operation, SC31-8567.
The APPN class of service consists of line and node characteristics. Line
characteristics include security, line speed, propagation delay and cost (in terms
of units of time and by byte). Node characteristics are the level of congestion
allowed and the desirability of the node for intermediate routing. This
desirability of a node for intermediate routing is called the route resistance.
The APPN COS also specifies which of four transmission priorities (low, medium,
high, network) a session is to take.
Note
Chapter 1. Introduction 15
network node server. At present only VTAM provides support for session
services extensions, although 2216 support has just been announced.
LFSID is not used by HPR. Instead, sessions are identified by the session
address in the FID5 header.
An NCE identifies an RTP endpoint within an HPR node. Exactly what functions
an NCE represents depends on the implementation. VTAM usually has one NCE
for all its LU-LU sessions, one for CP-CP sessions, and one for Route Setup
flows. Other NCEs may be defined for specific functions.
A TCID identifies a particular RTP pipe as seen by one of the RTP partners.
Each partner assigns a TCID to the connection which is subsequently used by
the other partner in the RTP header it sends. Thus an RTP pipe has two NCEs
(one at each end) and two TCIDs (one in each direction).
When CP-CP sessions have been established, topology updates and resource
registration flows take place. Once the network topology has been established
between network nodes and connections made to end nodes, LU-LU sessions
can be started. To establish an LU-LU session, there are four basic steps, as
follows:
1. The LU specifies the desired attributes for the session. It provides a mode,
which is resolved to an APPN COS by its CP (or, in some cases, its NN
server).
2. The destination LU (DLU) is located. Based on the information in its topology
and directory databases, the NN server issues one of the following:
• A directed search to the DLU′s NNS (if the NNS has knowledge of the
DLU location in its directory database and therefore needs only to verify
that it has not moved)
• A referred search (to search the central directory server′s database)
• A broadcast search (issued by either a central directory server, or a
network node server if there is no CDS in the network)
3. The network node server of the primary LU chooses the optimal route. The
topology database, the session COS and the APPN COS table are used to
calculate the route.
Note that, if the session path contains an IC-TG, the network node server of
the OLU (which is not necessarily the PLU) will calculate the route. This is
necessitated by the differences in the session setup flows for APPN and
subarea.
4. Finally, the session is established by forwarding the BIND from the PLU to
the SLU. The BIND contains a Route Selection Control Vector (RSCV) built
by the NN which calculated the route. The RSCV identifies the nodes and
TGs along the route from the PLU to the SLU. The RSCV allows the BIND to
flow on its chosen path, as each intermediate NN builds its routing tables to
match the LFSIDs to the session.
For HPR-capable networks, the session setup is the same as for basic APPN with
some exceptions. The network topology database is aware of which nodes are
Chapter 1. Introduction 17
RTP-capable, which are ANR-capable and which are neither. When the RSCV is
created, it contains the same information which is therefore available to
CP(PLU). Thus the PLU node can identify where on the session path an RTP
connection is possible. If a suitable RTP connection already exists (same APPN
COS and same HPR portion of the route as the new session), it is used for the
new session. Otherwise, a route setup message flows at the point where the
BIND would be sent in base APPN. The route setup protocol obtains ANR
forward and reverse labels for the route. The RTP connection is then
established and the BIND is sent on it.
Note that the VTAM systems described here are all V4R3. When the tests
described in this book were conducted, V4R3 was the current level of VTAM.
V4R4 was used in the second volume. The major differences between V4R3,
V4R4 and CS OS/390 Release 5 (the latest available level of VTAM) are in the
HPR and sysplex areas. In this volume, the differences between V4R3, V4R4 and
CS OS/390 R5 operation are minimal, and are pointed out where they occur.
Note: VTAM RAA is the CMC in this subarea network and thus owns all the SNA
network resources defined on the NCPs, 3174 controllers, CM/2 and CS/2
machines, and AS/400. VTAM RAS is only a data host.
Chapter 2. N e t w o r k Configuration 21
has been included in Appendix F, “VTAM Definitions with RAA as ICN” on
page 243, Appendix G, “VTAM Definitions with RAA as ICN and RAS as MDH”
on page 245, and Appendix H, “VTAM Definitions with RAA as ICN and RAS as
EN” on page 251.
Note
The migrated network we have shown here has both subarea and APPN
connections between the interchange node RAA and the migration data host
RAS. This is because our objective was to show every possible type of
connection and migration scenario. In real life, we do not recommend that
parallel APPN and subarea connections remain between any two VTAMs.
Either use VR-TG (in which case you have parallel APPN connections), or
migrate all FID4 links to FID2 (native APPN). 5.6, “Changing the Data Host
from MDH to EN” on page 108 and Chapter 7, “VR-Based TG
Implementation” on page 131 describe these two more desirable endpoints
of a migration.
Chapter 2. N e t w o r k Configuration 23
24 Subarea to APPN Migration: VTAM and APPN Implementation
Chapter 3. Planning for Migration
The planning of an APPN environment is the most important part of the whole
process of migrating towards an APPN network, whether you as a user have a
large or a small SNA network. Get this right and you will have few problems.
There are many tasks to be done on the drawing board before touching a single
parameter. Additionally, there are many changes to hardware and software,
which can be made before initializing APPN on VTAM nodes.
This chapter concentrates on these planning steps and gives the customer an
insight into the most important ones for subarea to APPN migration. The
scenarios described later in this book are for a small- to medium-sized network.
In spite of this, the reader will still find the scenarios relevant to any network.
We have two VTAMs and two NCPs where other networks may have dozens of
VTAMs and NCPs. The migration process will be fairly similar. Customers will
have to decide on various parameters based on their own networks and how
those networks are presently defined. We will give some guidance on how to
answer these questions. This chapter therefore outlines guidelines for different
customer situations and leads into the practical chapters where we have
migrated our own network.
Most customers will want to know how to incorporate their existing dependent
logical units, such as 3270s, into an APPN network. This is covered in this
planning chapter and also in subsequent chapters where the need arises. The
next stage of dependent LU support (DLUR/DLUS) is covered in SG24-5204, as is
migration to remote DLUR nodes and HPR routers such as the 3746-9X0, 2216
and 2210.
In this book we have also covered subnetting (using the extended and peripheral
border node functions) to show how to split networks into smaller, more
manageable networks with less broadcasting of topology exchanges. Customers
will also see how the border node function can be used to migrate from SNI
environments.
Chapter 9 to Chapter 12, although not based on testing in our limited laboratory
environment, give some guidance relevant to larger networks. They cover
issues such as searching, route selection, determining class of service and
APPN network management. The larger the network, the more significant these
topics become.
Larger SNA sites will want to reduce reliance on definitions but will be reluctant
at first to do this for a number of reasons. Amongst these are security and the
Note that VTAMs and NCPs not supporting APPN may participate in APPN
networks as pure subarea nodes, LEN nodes, or as components of composite
network nodes.
You should review all the available maintenance for VTAM, particularly as new
function is often added via the maintenance process in between VTAM releases.
Where such functions are referred to in the book, we quote the APARs that
introduced them. We also list the recent ones in Appendix K, “New Function
Added by APARs” on page 261.
Coding HOSTSA and NODETYPE=NN for the CMC and the backup CMC allows
both VTAMs to load, activate and own NCPs. If the backup VTAM was coded as
an EN (NODETYPE=EN), that VTAM could not participate in a network
takeover/giveback because the NODETYPE= start option cannot be changed
dynamically to convert the EN into an NN. The primary CMC VTAM will be a
CNN since it owns the network and its NCPs. After SSCP takeover, the backup
CMC becomes a CNN. The backup ICN will require FID4 connections to the
subarea network to allow it to take over NCPs. Refer to Chapter 8, “SSCP
Takeover in a Combined Subarea and APPN Network” on page 141 for more
information on APPN SSCP takeover.
Other VTAMs in the network would be designated as data hosts in the subarea
network (for example, they have channel-attached major node connections to
NCPs). These should initially be migrated to migration data hosts (MDHs) with
HOSTSA= and NODETYPE=EN specified in the ATCSTRxx member. These can
If you only have one VTAM in the network, make this an ICN as above. For all
intents and purposes, this is the CMC.
Another VTAM role to plan for is the designation of a central directory server.
Depending on the size of your network, you should have at least one CDS in
your network. The default registration value for VTAM applications is to register
with the NN server and a CDS. Using a CDS will reduce network broadcast
searches.
If you have an SNA network that consists of three or four remote sites, each of
which contain multiple VTAMs and NCPs, you may want to think about
configuring a CDS (or two) at each of these sites to reduce search traffic
between sites. It is common for the ICN to be given the role of central directory
server. Alternatively or in addition to using an alternate CDS, use subnetting.
See 3.1.16.1, “Setting up a Central Directory Server” on page 46 for more
information.
When designing your APPN network using your existing network diagrams, you
will want to decide the following:
• Which nodes to designate as NNs and which nodes to designate as ENs
• Which network connections can be converted to FID2 APPN connections and
which links must remain FID4 subarea links
In making these decisions, you will want to know the storage limitations of
various APPN platforms that comprise your network, the traffic flow patterns in
your network, and the traffic type. Understanding these issues will help you plan
for use of the border node function, plan adequately for the coding conventions
influencing network searches, and avoid an invalid network design.
For example, if you discover in your network analysis that you have existing
APPN subnets whose NNs have limited storage capacity (3174, for example), you
may be led to a design decision involving continued LEN connectivity or
peripheral boundary attachment until such a time as the capacity constraints can
be overcome with either a replacement platform or with an APPN subnet
redesign.
LEN or peripheral boundary connection will isolate the APPN subnet from the
greater VTAM/APPN network, thus respecting the capacity limitations of the NNs
in question. For more information and migration scenarios on APPN subnetwork
boundaries, see Chapter 6, “Border Node Support” on page 115.
A proper understanding of network attachments and traffic flows will make the
task of coding certain VTAM start options (SORDER, SSEARCH, and so on) and
VTAMLST members (ADJCLUST, ADJSSCP, and so on) more logical than
otherwise would be the case. For more information about the VTAM start
options and VTAMLST, see VTAM Resource Definition Reference, SC31-8565, and
VTAM Network Implementation Guide , SC31-8563.
Finally, your network design must be valid, allowing for all required
communication, thus avoiding the following common design errors:
Consider the situation in the upper diagram in Figure 3. Two VTAM/NCP sites
are connected via a null SNI network, which isolates their network topologies yet
permits any LU-LU sessions across the network boundary. If it is desired to
replace one of the NCPs by a 3746-950 as in the lower diagram, the following
considerations apply:
1. Both VTAMs must be converted to APPN because the 3746 does not support
subarea connections.
2. VTAM2 must become an NN because it owns an NCP.
3. VTAM1 must also become an NN because at the time of writing the 3746
does not support session services extensions. Thus if VTAM1 was an end
node it would have no suitable network node server.
4. Because VTAM1 and VTAM2 are network nodes with different NetIDs, there
must be an APPN border between them. This can be on either side of the
3746.
5. Because the 3746 does not presently support the border node function, such
an APPN border must be a peripheral border.
The result is that dependent LU sessions cannot traverse the APPN connection
between the two VTAMs. Even if DLUR is implemented on the 3746, not all
configurations will work. You must wait until the 3746 (or the 2216, for that
matter) supports the extended border node function, of which session services
extensions are a subset.
Stop Press
Support for the extended border node and session services extensions
functions has just been announced for the 2216. Planned availability is June
1998.
We have dealt earlier in this section with VTAM and NCP roles, but we have not
looked at other nodes which will participate in your APPN network. There are
many such devices that act as NNs, ENs or LEN nodes. These include AS/400,
RS/6000, PS/2 and many others. 3746, 2216, 2210 and 3174 nodes can participate
in APPN but cannot be configured as end nodes.
You may want to cluster ENs around non-VTAM NNs to reduce the number of
CP-CP sessions with VTAM and the registration traffic that accompanies these.
However, in doing so, you must take into account not only the capacity
limitations of various NN platforms, but also their usefulness as NNs in a session
setup and in a failure/recovery situation. Remember that a VTAM EN must have
another VTAM as its NN server because, at the time of writing, only VTAM
supports session services extensions.
In the case where a VTAM fails and another VTAM takes over its NCPs, you
should be aware that:
• All CP-CP sessions to the failed VTAM will be terminated, even though the
NCP and cross-domain LU-LU sessions remain active.
• A new XID exchange will take place to establish the new relationship
between the connection partners.
• The CP name presented to the adjacent node will change.
Therefore, for takeover to function smoothly attached nodes must be able to:
• Tolerate receipt of XIDs on an existing connection
• Tolerate CP name change when the CP-CP sessions are reestablished
Some nodes do not support these functions (APPN options 1001 and 1004).
Therefore, you may want to upgrade them or replace them with platforms that do
Although the 3174 can only be defined as an NN, it is not a good choice for an
NN. This is due to its traditional boundary node function and the fact that it has
minimal storage capacity for topology databases. It is recommended that it
either be left connected via LEN, or isolated from the backbone using VTAM′ s
border node function.
When migrating to APPN, you will need some conventions for new resources
such as CP names and possibly additional network names (see 3.1.17, “Border
Node” on page 49 and Chapter 6, “Border Node Support” on page 115 for more
information on subnetting and APPN borders).
It is quite possible that your network will have APPN-capable resources with
duplicate CP names, which are currently attached as LEN nodes. This must be
resolved prior to nodes becoming APPN nodes. Specifying CONNTYPE=LEN for
a connection causes VTAM to avoid checking for duplicate CP names, except
when CPNAME is coded on a switched PU. At the same time, it forces the
connection to remain LEN even though both partners are APPN-capable.
Also, when converting a LEN to an APPN node, ensure that the adjacent link
station name (VTAM-defined PU name) for an adjacent node is different from the
CP name of that node; the PU name and the CP name are now in the same
name space and will be considered duplicate names if they are the same.
CP-CP sessions will not be set up and a sense code will be issued by VTAM.
* RAAAN D NET,ID=RAATRPU2,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = RAATRPU2 , TYPE = PU_T2.1
IST486I STATUS= ACTIV , DESIRED STATE= ACTIV
IST1043I CP NAME = RAATRPU2, CP NETID = USIBMRA , DYNAMIC LU = NO
IST1105I RESOURCE STATUS TGN CP-CP TG CHARACTERISTICS
IST1106I RAATRPU2 AC/R 21 YES2 982D000000000000000001710080808
IST1482I HPR = NO - OVERRIDE = N/A - CONNECTION = NO
IST136I SWITCHED SNA MAJOR NODE = RAASWMN2
IST081I LINE NAME = J0005021, LINE GROUP = EG05L01 , MAJNOD = RA5NCPX
IST654I I/O TRACE = OFF, BUFFER TRACE = OFF
IST1500I STATE TRACE = OFF
IST355I LOGICAL UNITS:
IST080I RAATLU22 ACTIV RAATLU23 ACTIV RAATLU24 ACTIV
IST080I RAATLU25 ACTIV
3
IST314I END
Figure 4. Displays and Sense Codes for RAATRPU2
Figure 4 shows the displays observed in VTAM. What happened here was that
the APPN connection was established successfully (after XID3 exchange), but
then the PC tried to establish CP-CP sessions with VTAM. The NCP sent in
BFINIT, but VTAM rejected this with a sense code 1 which means “BFINIT
invalid for resource type”. The BFINIT has the OLU name RAATRPU2 on it, but
VTAM regards it as a PU name.
In this case there is no match between the CP name coded in the switched
major node and the actual CP name presented on the XID3 from CM/2. VTAM
set up the connection because dynamic PU definition was in effect, but no CP-CP
sessions were possible because the LU name of the CP was still the same as
VTAM′s link station name.
Now the definitions are correct and the names are unique. Figure 5 on page 33
shows the VTAM display of the link station RAATRPU2. Now RAATRCP2 is seen
in the list of sessions 1 on this link station, so we know that the PC′s CP name
is a valid LU name. The display of VTAM′s topology database shows at 2 that
there are indeed CP-CP sessions between the two nodes. The following results
were observed:
* RAAAN D NET,TOPO,LIST=EN
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.WTR05115 EN *NA* *NA* YES *NA*
IST1296I USIBMRA.CP05147 EN *NA* *NA* YES *NA*
IST1296I USIBMRA.RAATRCP2 EN *NA* *NA* YES2*NA*
IST1296I USIBMRA.CP05137 EN *NA* *NA* YES *NA*
IST1296I USIBMRA.RAATRPU2 EN *NA* *NA* NO 3*NA*
IST314I END
Figure 5. Displays of RAATRPU2 with C P N A M E = R A A T R C P 2
Note the “----Y” in the entry for RAATRCP2 1, meaning that the resource has
been created dynamically. All independent LUs in session with VTAM′s own
resources are represented by CDRSCs. Although the new CP name is defined
as a CP name in the switched major node, it is not defined as an LU anywhere.
Therefore, the CDRSC is created dynamically when the BIND arrives for the first
CP-CP session.
Note also that the previous CP name 3 of the PC is still in the topology
database since the default time for resources to remain in the topology database
is 15 days. You can manually delete this entry using the following command:
F NET,TOPO,ID=RAATRPU2,TYPE=FORCE
Note: You may be asking why ENs are shown in this display, when only network
nodes are represented in the topology database. In fact, the D NET,TOPO
command causes VTAM to display its local topology as well as the network
topology. Thus attached end nodes are displayed as well as attached network
nodes.
The lesson learned here is to review your naming conventions for CP names and
PU names.
For network names (NetIDs), you can usually use your existing names. If you
decide to subnet, you may require more network names. Depending on your
strategy you may have the same NetID for different APPN networks or use a
different one for each. See 3.1.17, “Border Node” on page 49 for more
information.
3.1.4 IBM 3746 Model 900 and Model 950 APPN Migration
If you use 3745-17A, 3745-21A, 3745-31A, 3745-41A or 3745-61A controllers, you
may also have the Model 900 expansion unit attached. This can offload DLC
functions in separate processors for SDLC, token-ring, frame relay, X.25 and
ESCON connections. The 3746-950 is a stand-alone version of the 3746-900, from
which an upgrade path is available. These controllers provide an excellent
migration path to APPN.
For information on APPN migration using the 3746 Model 900 and 3746 Model
950, please refer to the HPR and DLUR Implementation volume of this book,
SG24-5204.
Note:
2If you do not want dynamic builds of PUs and dependent LUs at this
time, replace the * CMNT * on this line. The default exit has no comment.
Reassemble and link-edit the exit, then re-install according to your site′s normal
procedures. The MODIFY EXIT command enables you to activate, deactivate, or
replace the exit without recycling VTAM.
All connection status records will be recorded in the CSDATA data set for later
processing. Refer to VTAM Customization , LY43-0110 for a description of this
exit and the IBM-supplied default which can also be found in
SYS1.ASAMPLIB(ISTEXCCS). Users must be licensed for VTAM to obtain this
manual.
VTAM provides three methods of defining switched adjacent link stations, two
dynamic and one static:
• Configuration services XID exit ISTEXCCS
In conjunction with the model terminal support major node, this exit can be
used to define dynamically PUs and LUs for switched incoming connections.
This also includes leased connections that appear to VTAM as switched
connections, such as token-ring devices and DLUR resources. The user
must define the exit and a model major node. The model major node is
coded using VBUILD TYPE=MODEL. The IBM-supplied default exit is coded
with the dynamic build function activated.
• DYNPU definition keyword.
This method is also available only for switched incoming connections, but
this time there is no provision for dependent LU definition. The DYNPU
If dependent LUs are required, then the recommended method to use is the
configuration services XID exit, since DYNPU=YES causes only a PU (no LUs) to
be created. The ISTEXCCS exit gives control over PU and LU names and allows
security control in conjunction with the session management exit.
To give you a quick insight into the way the exit works, we have included a
description in Appendix J, “Configuration Services Exit” on page 257.
Note
It is not recommended that you change this table or add to it. If you do
decide to do this, you will have to update every APPNCOS table in every
APPN node through which your session may pass.
You may want to put an entry for your APPN COS table in the ATCCONnn
configuration list so that VTAM activates it at initialization. However, if VTAM is
started as an APPN NN or EN, it will activate the one called COSAPPN
automatically if it is found in VTAMLST.
There are two new VTAM start options associated with APPN class of service,
namely APPNCOS and ISTCOSDF. APPNCOS specifies an APPN class of service
which VTAM will use for a session under certain circumstances (see Figure 8 on
page 39). During initial testing APPNCOS should default to NONE. This will
allow you to resolve any APPN COS resolution problems, since sessions with an
unknown APPN COS will fail. After testing, change the default to #CONNECT to
allow unresolved APPNCOS sessions to use a medium-priority class of service
with relatively undemanding route requirements.
ISTCOSDF tells VTAM which types of resource can use the default mode entry
ISTCOSDF, and is invoked if no mode name is available for a session. This
could happen in an APPN-to-subarea session setup situation, as described in
Chapter 10, “Logmode and COS Resolution in a Mixed Network” on page 167.
For information on how VTAM selects an APPNCOS entry, see 3.1.9, “ISTINCLM
and User-Defined Logmode Table Requirement” on page 38 and Figure 8 on
page 39.
Note
The TG profile table is useful in defining profiles which can be pointed to by the
TGP= parameter on APPN connections. Although all of the parameters which
make up a transmission group profile table entry can be specified on each
connection, you may feel it is more easily managed by having a single table.
You can add entries to this table depending on your network setup. The table
should be copied from SYS1.ASAMPLIB (IBMTGPS) to your VTAMLST data set.
You should put an entry for IBMTGPS in your ATCCONnn configuration list so
that VTAM activates this table. This table is not assembled and link-edited.
If you activate a new TGP definition, it is used for subsequent link activations but
will not update existing TG characteristics unless you issue a MODIFY TGP
command for the existing TG.
There are different ways to approach this. The IBM-supplied table ISTINCLM has
been updated with the APPNCOS= parameter for every entry and therefore will
not pose a problem. Although the new version is supplied in SYS1.VTAMLIB, if
you do not use this data set in your VTAMLIB concatenation you should check
that the new version of ISTINCLM will be picked up by the VTAM start procedure.
For user-defined logmode tables, you can do one or more of the following three
things:
1. Update all your logmode entries with an APPNCOS operand, which equates
to one of the entries in the APPNCOS table COSAPPN. The value chosen
will depend on the session path selection criteria.
Note
CS OS/390 Development strongly recommends that you consolidate all
your user-defined mode tables into a single table.
The APPNCOS is selected in a different manner from subarea COS. In APPN the
COS is chosen at the OLU end of the session, whereas in a subarea network the
COS is chosen at the SLU end.
VTAM will select an APPNCOS for a session in the following way if there is no
SATOAPPN COS mapping table:
If COS mapping tables are used, then the input COS (subarea or APPN) is used
to determine the output COS (APPN or subarea, respectively). The first COS
resolved by VTAM is always a subarea COS, and is always chosen by looking up
mode tables. Subsequent COS resolution is done by using the mapping tables if
they exist.
For a more detailed discussion of COS selection and mapping, please refer to
Chapter 10, “Logmode and COS Resolution in a Mixed Network” on page 167.
MPC is available on the more recent types of CTC connection, as far back as the
IBM 3088 and the VM virtual CTC links. However, it may not be supported on
older or non-IBM hardware and you will need to check this with the appropriate
supplier. If MPC is not supported then VR-TG is the only way to implement
APPN over the CTC connection. VTAM V4R2 or later is required to run APPN
over multipath channel connections.
*******************************************************************
* *
* CA MAJORNODE FOR RAA(10) TO RAS(28) - FOR FID4 MPC CONNECTION *
* *
*******************************************************************
RAACRAS VBUILD TYPE=CA
RAAGMP1 GROUP LNCTL=MPC,ISTATUS=ACTIVE,MAXBFRU=3
RAALMP1 LINE READ=(C12), *
WRITE=(C14)
RAAPMP1 PU PUTYPE=4,TGN=1
*******************************************************************
* *
* CA MAJORNODE FOR RAS(28) TO RAA(10) - FOR FID4 MPC CONNECTION *
* *
*******************************************************************
RASCRAS VBUILD TYPE=CA
RASGMP1 GROUP LNCTL=MPC,ISTATUS=ACTIVE,MAXBFRU=3
RASLMP1 LINE READ=(C14), *
WRITE=(C12)
RASPMP1 PU PUTYPE=4,TGN=1
These are examples and users should arrange the channel addresses with their
MVS systems programmer prior to implementing. See VTAM Resource
Definition Reference , SC31-8565.
This is all you need for the MPC for subarea connectivity, although for better
availability and performance you may require multiple READ and WRITE
subchannels. See the chapter on tuning I/O for MPC connections in the Network
Implementation Guide , SC31-8563.
This is done by changing the CA major node for MPC to a local SNA major node
with a TRLE operand pointing at a transport resource list element name. The
At the same time create a transport resource list (TRL) definition specifying
L N C T L = M P C . The TRL element (TRLE) is not a resource but describes the
connectivity characteristics of the multipath channel connection. VTAM places
all TRLE definitions in a single major node labelled ISTTRL, which cannot be
deactivated. If you need to change the definition while it is active, use the
command V NET,ACT,ID=trl major node name,UPDATE=ALL.
These LOCAL and TRL definitions must be created at both VTAM nodes and
must match READ and WRITE channels in the same way as the CA major node
for MPC. The TRL major node must be activated before the PUs that reference
it; otherwise the PU activation will fail. Therefore, the local SNA major node for
an ANNC connection must be defined in the ATCCONnn member after the TRL
major node.
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAA (SA10) TO RAS (SA28) *
****************************************************************
RAAAHHC VBUILD TYPE=TRL
*
RAAAHHCT1 TRLE LNCTL=MPC,
READ=(C14),2
WRITE=(C12)36
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAA (SA10) TO RAS (SA28) *
****************************************************************
RAAAHHC VBUILD TYPE=LOCAL
RAAAHHC1 PU TRLE=RAAAHHCT,1
XID=YES,4
CONNTYPE=APPN,
CPCP=YES,
TGP=CHANNEL,5
ISTATUS=ACTIVE
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAS (SA28) TO RAA (SA10) *
****************************************************************
RASAHHC VBUILD TYPE=LOCAL
RASAHHC1 PU TRLE=RASAHHCT,1
XID=YES,4
CONNTYPE=APPN,
CPCP=YES,
TGP=CHANNEL,5
ISTATUS=ACTIVE
Notes:
1The label on the TRLE definition must match the TRLE= operand in the ANNC
local major node.
2The READ address C14 in RAA must relate to the WRITE address on the RAS
side. In our configuration each subchannel has the same address on each side.
3The WRITE address C12 in RAA must relate to the READ address on the RAS
side.
4Make sure that XID=YES is defined; otherwise, this is seen as a type 2.0
node and activation will fail.
Once a VTAM host is made APPN-capable, the default start options will
automatically try type 2.1 connections as APPN connections. If your naming
convention for CP names and PU names has not been reviewed, there may be a
conflict between the VTAM PU name, the VTAM-defined CPNAME and the
attached node′s CP name. If there is a conflict, you may find that these nodes
will still be connected as LEN nodes. See 3.1.3, “Naming Conventions” on
page 30 for more information.
You can refer to the scenarios in Chapter 4, “Migrating a CMC Host to ICN” on
page 61 for a practical demonstration of what happens to different type 2.1
devices with different configurations. The node type of each adjacent type 2.1
node can be specified on the PU by NN=YES or NO. If this differs from the
attached node′s actual role, the connection will fail. If not coded, the node type
will be determined in the XID3 when the connection is activated; this is the
recommended way.
These NCP values should already be in place if you are using LEN connections,
but migration to full APPN is likely to result in more independent LUs
establishing more sessions with your VTAMs.
If you set up CP-CP sessions between VTAM and any adjacent node, the CP
name is automatically defined as a CDRSC by VTAM regardless of your other
definitions or start options. However, if ILUs other than the adjacent CP attempt
to set up sessions across the connection, you will have to ensure that dynamic
LU definition is in effect. This can be achieved in one of three ways:
• Code DYNLU=YES as a start option (the easiest and the recommended
option).
• Code DYNLU=YES on the ADJCP statement (if using static CP definition).
• Code DYNLU=YES on each PU (link station) definition for the adjacent node.
Unless these independent LUs are defined (statically or via DYNLU) you will see
session setup failures such as:
After this error, we created and activated the following adjacent CP major node:
*******************************************************************
* *
* ADJCP TABLE FOR RAA *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
* RAS SUBAREA 28 IS THE ONLY OTHER VTAM IN THE NETWORK *
* CP05147 IS AN APPN NODE RUNNING COMMUNICATION SERVER *
* *
*******************************************************************
RAAADJCP VBUILD TYPE=ADJCP
RAS ADJCP NETID=USIBMRA,DYNLU=YES
CP05147 ADJCP NETID=USIBMRA,DYNLU=YES
The display of the link station then confirmed 1 that dynamic LU definition was
in effect:
Only VTAM supports the CDS function, although most other products will make
use of a CDS if one exists in the network. In general, you would choose your
CMC host VTAM(s) to be the CDS(s).
LEN nodes cannot perform registration because they have no CP-CP sessions
with their network node server. Any target resources on a LEN node must be
predefined somewhere, normally on their NN server.
We found that the default registration values were sufficient in the laboratory
network. You may want to review their values.
Set up a preferred NN server list in each VTAM EN or MDH. If you do not specify
this list, the VTAM EN or MDH will set up CP-CP sessions with the first network
node it finds. Apart from the unpredictability of this, if VTAM chooses a
channel-attached 3746 or 3174 as its NN server your network will not function.
Dependent LU sessions require the APPN session services extensions option in
both EN and NN server; only VTAM currently supports the option.
You can control (to some extent, as described above) which NN is the preferred
NN server and define secondary NN servers. For example, you may want to put
your takeover ICN host node as the secondary preferred NN server.
To set up an NN server list in an EN, add a network node server list major node
to VTAMLST. An example is shown below. Activate your network node server
list by adding it to your ATCCONxx configuration list and restarting VTAM, or
activate it using the following command:
V NET,ACT,ID=network node server list
*******************************************************************
* *
* NETWORK NODE SERVER LIST MAJOR NODE FOR RAS *
* USES RAA AS ITS NETWORK NODE SERVER *
* THIS IS SUBAREA 28 PRIOR TO MIGRATION FOR RED BOOK *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
RASNNSVR VBUILD TYPE=NETSRVR,
ORDER=FIRST 1
RAA 2 NETSRVR NETID=USIBMRA,
SLUINIT=REQ
If you wish to change your list dynamically, this can be done by issuing
VARY,NET,ACT,ID=netsrvr list name after updating the table. Then you must issue
VARY NET,TERM on the CP-CP sessions between this EN and the existing NN server
to force VTAM to use the new list.
Note: Issuing a VARY TERM command to terminate the CP-CP sessions might
cause later session establishment requests to fail. To resolve the problem,
reactivate the CP-CP sessions or else reactivate the link with CPCP=NO so that
it will no longer be used in directed search routing. Refer to VTAM Operation ,
SC31-8567.
The first thing to remember is that there are two types of border (peripheral and
extended) and two types of border node (peripheral and extended). An extended
border requires two extended border nodes (EBNs); both EBNs share the
management of the border, and full support is available for the establishment of
any type of session. A peripheral border is managed by just one border node,
which can be an EBN or a peripheral border node (PBN). There are restrictions
on the establishment of sessions across a peripheral border, and even more
restrictions if it is managed by a PBN.
A peripheral border is formed between two NNs, where one (the border
manager) pretends to be an EN and the other takes no part in controlling the
border. Since an EN is allowed to have a different NetID from its NN server, this
works quite well. However:
• A peripheral border must be the first and/or the last border on a session
path. If a session crosses more than two network borders, all the ones in
the middle must be extended borders.
• A peripheral border managed by a PBN can be the only border on a session
path. Only two subnetworks can be connected in this manner.
• Because a peripheral border relies on one node having an EN appearance,
dependent LU sessions crossing such a border need session services
extensions support in both nodes. This means that both nodes must be
VTAM, so you might as well have an extended border with more function.
An extended border is formed between two NNs where both support the APPN
extended border node function, option 1016.
Secondly, if this is an external SNI link to a trading partner, has the partner
migrated to APPN? If not, you must leave the connection as SNI. If the partner
does support APPN, do you want to change the SNI link to an APPN border node
connection? Remember that the APPN node on the partner′s side must also be
VTAM (or NCP, but not a 3746 or 2216, at the time of writing) if you want any
dependent LU sessions across the border.
Your SNI setup may include the use of VTAM′s session management exit and be
complicated by specific SNI GWNAU statements for security purposes.
Converting this to APPN may require a re-design of the code and the use of the
APPN directory services management exit (DSME) to enforce the same level of
security that you have already. The question is, can an APPN border do what
you already do with SNI?
To subnet you must decide where you want your borders to be. You can choose
a different NetID or keep the same NetID throughout your network.
If your CMC is to be converted into an EBN, make sure that your takeover VTAM
ICN is also capable of EBN support by coding the same parameters in that
VTAM′s start options. This will allow it to take over border node connections
during takeover processing.
To turn on border node support in VTAM, specify BN=YES in the VTAM start
options of your chosen NN(s). There are a number of controls available to
customize your border node support:
1. VTAM start options
• BN=YES turns on border node functions for this VTAM NN.
See Chapter 6, “Border Node Support” on page 115 for some migration border
node scenarios.
For example, you may have two ENs on a token-ring that wish to communicate
with each other. Without a connection network, the connection between them
would have to be predefined on at least one of them; otherwise all sessions
would be routed via their NN server. With a connection network they can
communicate directly without this extra definition.
In both cases you need DYNPU=YES (the default). A transmission group profile,
TGP, can be specified on the NTRI physical LINE and the XCA major node PORT
definition statements.
If your installation does not allow dynamic adjacent CP definition, you can
predefine the virtual node as an adjacent CP by coding VN=YES on the ADJCP
definition.
If a VTAM NN owns multiple TICs that are defined on the same connection
network, you may find that sessions from other nodes are distributed throughout
the TICs. Thus (for example) an end node with four parallel sessions to a VTAM
application may use four TICs, and therefore four logical lines. This obviously
has an impact on NCP storage. One way round this, if the VTAM NN is the NN
server of the nodes which are setting up these sessions, is not to define the NN
on the connection network. The links used for the CP-CP sessions have to be
predefined in any case, and the NN can still calculate routes between ENs across
the connection network.
A VTAM interchange node provides directory services for both APPN and
subarea networks. A VTAM migration data host can receive a request for a
search from a resource within its domain, and this request can be forwarded into
either the APPN or the subarea network. Thus each node type may have to
decide whether a given search request is forwarded into subarea or APPN, or
both, and in which order. VTAM provides options for you to use to control this
process.
*******************************************************************
* *
* ADJSSCP TABLE FOR RAA *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
* RAS SUBAREA 28 IS THE ONLY OTHER VTAM IN THE NETWORK *
* *
*******************************************************************
VBUILD TYPE=ADJSSCP
RAK ADJCDRM
ISTAPNCP ADJCDRM
RAS ADJCDRM
* RAAAN D NET,ADJSSCPS,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = ADJACENT SSCP TABLE
IST623I DEFAULT ADJACENT SSCP TABLE
IST624I RAK
IST624I ISTAPNCP
IST624I RAS
IST1454I 3 RESOURCE(S) DISPLAYED
IST314I END
Figure 19. ADJSSCP Table Display with ISTAPNCP Entry
With search reduction, the user can specify the number of requests that will be
denied once a search target has been determined to be unreachable
(SRCOUNT), and/or specify a timer value (SRTIMER) reflecting a period of time
during which VTAM will deny session requests for that resource. If the network
had just been started and all of the 10,000 LUs tried to log on to a nonexistent
resource, carefully chosen search reduction parameters could make VTAM send
out just one search instead of 10,000. The three new parameters are as follows:
1. SRCHRED. This enables search reduction. The default is OFF.
2. SRCOUNT. The default for this is 10 (VTAM suppresses ten searches before
trying to locate the target again). For larger networks, this can be increased
to 1000 but will vary depending on your individual network configuration.
This is a parameter to watch after APPN migration.
3. SRTIMER. This defaults to 30 seconds (VTAM suppresses searches for 30
seconds before trying to locate the target again). As for SRCOUNT, the
optimum value varies depending on your network size.
All three search reduction options can be changed dynamically using the F
netproc,VTAMOPTS command.
Using the 3746 Model 9X0 3.1.4, “IBM 3746 Model 900 and
Model 950 APPN Migration” on
page 34
Multipath channel (MPC) conversion prior to APPN migration 3.1.10, “Multipath Channel
Conversion” on page 39
APPN node roles and connectivity 3.1.12, “APPN Node Roles and
Connectivity” on page 43
SNI and border node considerations 3.1.17.1, “SNI and Border Node
Considerations” on page 50
Once all the planning for the design of the mixed APPN and subarea network
has been completed, we can begin the migration.
The test network has been described in previous chapters, but for convenience it
has been drawn again in Figure 20. VTAM RAA, which is also the CMC host in
the subarea network, will be converted to an interchange node. By creating an
interchange node, we introduce the option of establishing an APPN connection
with any APPN-capable node while maintaining the current subarea
configuration. An ICN is necessary in the network, since only an ICN can load
and activate an NCP and also establish SNI links to external subarea networks.
The CMC host must be migrated to an ICN before migrating any data host to an
MDH or an EN.
Under a subarea network, the CMC host being the focal point of the network is
normally the target adjacent SSCP for network searches. On the same grounds,
the ICN is the focal point in an APPN network and thus should be the central
directory server for an APPN network. Thus, the VTAM RAA host will also
provide the central directory server function for our network.
As mentioned before, we make the RAA node APPN-capable but it will still use
subarea protocols for session establishment in the subarea part of the network.
APPNCOS=#CONNECT 8
CDSERVR=YES, 3
CONNTYPE=LEN, 2
CPCP=YES, 4
DIRSIZE=0, 1 10 DEFAULT
DIRTIME=8D, 1 DEFAULT
DYNADJCP=YES, 1 9 DEFAULT
HPR=NONE, 5
HOSTSA=10, 6
INITDB=ALL, 1 DEFAULT
NODETYPE=NN, 6
NUMTREES=100, 1 DEFAULT
RESUSAGE=100, 1 DEFAULT
ROUTERES=128, 1 DEFAULT
SSCPNAME=RAA, 7 SSCPNAME/CPNAME
SORDER=SUBAREA, 11
SSEARCH=YES, 1 DEFAULT
Notes:
1These start options will default to these values if not coded. In our test, we
coded them for clarity.
5HPR=NONE must be coded if you do not want VTAM to provide HPR support.
HPR=RTP is the default for all VTAM APPN nodes from V4R4 onwards.
HPR=NONE is usually safe and will make problem determination easier in the
initial stages of migration. However, it will prevent the establishment of HPDT
connections such as XCF (sysplex Cross-System Coupling Facility), and MPC to a
2216. We coded HPR=NONE initially, but a better plan may be to let HPR=RTP
default and override it individually on non-HPDT connections until you are happy
that the connections work correctly.
6Both HOSTSA and NODETYPE=NN start options are required for VTAM to act
as an interchange node.
8Specify a value for APPNCOS, which will be used as a default APPN class of
service if a requested class of service cannot be found in the topology and
routing services class-of-service database. APPNCOS=NONE is the default
value. We recommend that you run with APPNCOS=NONE until problems are
encountered; then dynamically change to APPNCOS=#CONNECT until problems
are resolved, then switch back to NONE. Otherwise, you might end up routing all
sorts of unknown session traffic over the same links and performance will be
affected.
11The SORDER start option controls the order in which the APPN and subarea
networks are searched when a search request for an unknown LU is received at
this node from the subarea network. SORDER=APPN (the default value)
specifies that the APPN network is to be searched first (or almost first - see
Chapter 9, “Searching and Routing in a Mixed Subarea/APPN Network” on
page 151). Specify SORDER=SUBAREA or ADJSSCP initially, to ensure that the
subarea network is searched early on in the algorithm. This will ensure that the
operation is as close as possible to the subarea case, in the early stages of
migration. If you code SORDER=ADJSSCP you must also modify your adjacent
SSCP tables, as described in Chapter 9.
The first time VTAM is started with these new data sets, and the INITDB start
option is set to ALL or TOPO, the following VTAM message will be received:
IST1288I TOPOLOGY DATASET RETRIEVAL WAS NOT SUCCESSFUL, CODE = 9
The only entries that VTAM saves in the directory checkpoint data set are
dynamic entries. These are entries which have been made as a result of a
successful search, whether from this NN (which saves the target LU location) or
to this NN (which saves the origin LU location). Directory entries registered to a
CDS (but not an NNS) are also regarded as dynamic.
Note: If you only have one NN in your network, there is no need to checkpoint
as there will be no network searches.
By default, a blank subarea COS (and therefore a low transmission priority) will
be used for CP-CP sessions that traverse subarea VRs. This applies also to
DLUR/DLUS control sessions. Therefore, we recommend that:
A copy of the default APPN COS table supplied by IBM is shown in VTAM
Resource Definition Reference , SC31-8565.
The second phase is to migrate each device, individually, from LEN to APPN
connection by specifying CONNTYPE=APPN on the link station (PU) definition of
the device on the host. You could opt to merge the two phases together,
depending on the size of your network, by specifying CONNTYPE=APPN as a
VTAM start option. The CONNTYPE value on the PU statement will default to
APPN. This will result in all of the APPN-capable devices behaving as APPN
nodes in one step instead of individually. We advise that you proceed with care.
Notes:
3Shows that this APPN node supports CP-CP sessions with adjacent nodes.
We issue the DISPLAY MAJNODES command, with the result seen in Figure 23
on page 67.
Note that the major node ISTRTPMN for RTP is not part of the above list. This is
because, in VTAM V4R3, an ICN could not act as an RTP endpoint. This
restriction was removed in VTAM V4R4.
* RAAAN D NET,TOPO
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1306I LAST CHECKPOINT ADJ NN EN SERVED EN CDSERVR ICN BN
IST1307I NONE 1 0 1 0 0 1 1 0
IST314I END
Figure 24. Display of Topology Summary
* RAAAN D NET,TOPO,LIST=CDSERVR
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST314I END
* RAAAN D NET,TOPO,LIST=ICN
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST314I END
Figure 25. Display of Topology with L I S T = N N , LIST=CDSERVR, L I S T = I C N
We establish TSO sessions from a CM/2 node defined as an EN with RAA as its
NNS, as well as from a CM/2 node defined as a LEN node. Dependent LU-LU
sessions are established, but no CP-CP sessions are established with VTAM by
either node. The end node is capable of establishing CP-CP sessions, but
because VTAM has CONNTYPE=LEN specified it declares at connection setup
time that such sessions are not acceptable.
* RAAAN D NET,ID=RAATRPU4,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = RAATRPU4 , TYPE = PU_T2.1
IST486I STATUS= ACTIV , DESIRED STATE= ACTIV
IST1043I CP NAME = CP05137 , CP NETID = USIBMRA , DYNAMIC LU = NO
IST136I SWITCHED SNA MAJOR NODE = RAASWMN4
IST081I LINE NAME = J0005025, LINE GROUP = EG05L01 , MAJNOD = RA5NCPX
IST654I I/O TRACE = OFF, BUFFER TRACE = OFF
IST1500I STATE TRACE = OFF
IST355I LOGICAL UNITS:
IST080I RAATLU42 ACTIV RAATLU43 ACTIV RAATLU44 ACT/S
IST080I RAATLU45 ACTIV
IST314I END
Figure 26. Display of CM/2 PU
The display of the 3174 configured as an APPN NN (PU name P3174C1) is shown
in Figure 28. Once again, VTAM perceives this as a LEN connection.
* RAAAN D NET,ID=P3174C1,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = P3174C1 , TYPE = PU_T2.1
IST486I STATUS= ACTIV , DESIRED STATE= ACTIV
IST1043I CP NAME = CP31742 , CP NETID = USIBMRA , DYNAMIC LU = YES
IST081I LINE NAME = L3174C1 , LINE GROUP = RA5GSFL0, MAJNOD = RA5NCPX
IST654I I/O TRACE = OFF, BUFFER TRACE = OFF
IST1500I STATE TRACE = OFF
IST355I LOGICAL UNITS:
IST080I RA317411 ACTIV RA317C12 ACTIV RA317C13 ACTIV
IST080I RA317C14 ACTIV
IST314I END
Figure 28. Display of the 3174 Link Station/PU
4.2.3 3174 as an NN
In our test environment, the 3174-11L (PU name RAALCLP) and the remote
token-ring attached 3174-63R (PU name RAATRPU1) are left as type 2.0 devices.
The remote SDLC-attached 3174-11R (link station name P3174C1) is configured
as an APPN network node, and VTAM is told to treat the connection as APPN.
Figure 29 on page 70 shows the network diagram with the 3174s and their PU
names and CP names, where applicable, and can be used as a reference.
The APPN function in the 3174 has been enabled by responding to Question 510
with 1, as shown in Figure 30. Question 511 defines the 3174′s CP name.
Question 512 allows you to define a connection network on this 3174, so that
nodes on the LAN need not define the link to the 3174 unless it is their NN
server.
Figure 30. Configuration Panel on 3174-11R
Session Setup: When the NCP major node is recycled to effect the change
(changing XID= requires an NCP generation), the following messages are
issued to indicate that an APPN connection and CP-CP sessions have been
established:
Note: This group of messages will be issued for all APPN connections
established with CP-CP sessions. If no CP-CP sessions are established, then
only message IST1086I will be issued.
Various VTAM displays can be used to verify whether CP-CP sessions have been
established. A display of the CP name CP31742 in Figure 33 on page 72 shows
a number of interesting points.
The two CP-CP sessions are using the adjacent link station P3174C1 10, which
has the same name as the PU. However, the ALSLIST 7, which is the list of
link stations VTAM will use to find this resource if it receives a session request,
does not include P3174C1. This is because CP31742 is an APPN node and must
be found using APPN mechanisms, not by a predefined connection. The notation
ISTAPNPU 7 signifies to VTAM that it must use APPN to find the location of
CP31742.
As well as being seen as a CDRSC / ILU to the subarea side of VTAM, CP31742
is also in the APPN directory 11. It is known there as a “dynamic NN” entry
12, since it was neither predefined in VTAM nor registered with it. Its location
has been discovered dynamically.
* RAAAN D NET,ID=CP31742,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = USIBMRA.CP31742 , TYPE = ADJACENT CP 1
IST486I STATUS= ACT/S----Y, DESIRED STATE= ACTIV 2
IST1447I REGISTRATION TYPE = NO 3
IST977I MDLTAB=***NA*** ASLTAB=***NA***
IST1333I ADJLIST = ***NA***
IST861I MODETAB=***NA*** USSTAB=***NA*** LOGTAB=***NA***
IST934I DLOGMOD=CPSVCMG USS LANGTAB=***NA*** 4
IST597I CAPABILITY-PLU ENABLED ,SLU ENABLED ,SESSION LIMIT NONE
IST231I CDRSC MAJOR NODE = ISTCDRDY 5
IST1184I CPNAME = USIBMRA.CP31742 - NETSRVR = ***NA*** 6
IST1044I ALSLIST = ISTAPNPU 7
IST082I DEVTYPE = INDEPENDENT LU / CDRSC 8
IST654I I/O TRACE = OFF, BUFFER TRACE = OFF
IST1500I STATE TRACE = OFF
IST171I ACTIVE SESSIONS = 0000000002, SESSION REQUESTS = 0000000000 9
IST206I SESSIONS:
IST1081I ADJACENT LINK STATION = P3174C1 10
IST634I NAME STATUS SID SEND RECV VR TP NETID
IST635I RAA ACTIV/CP-S EE4B79FF13F208DF 0018 0001 0 0 USIBMRA 9
IST635I RAA ACTIV/CP-P F7FF6164D39CEBE3 0001 002C 0 0 USIBMRA
IST924I -------------------------------------------------------------
IST075I NAME = USIBMRA.CP31742 , TYPE = DIRECTORY ENTRY 11
IST1186I DIRECTORY ENTRY = DYNAMIC NN 12
IST1184I CPNAME = USIBMRA.CP31742 - NETSRVR = ***NA*** 6
IST314I END
Figure 33. Display of 3174 CPNAME CP31742
Figure 34 on page 73 also shows that HPR is not active 2 and that the 3174′ s
CP is in session 3.
Figure 35, Figure 36 on page 74 and Figure 37 on page 74 show the NetView
Session Monitor displays of the 3174 CP-CP sessions.
NLDM.SESS PAGE 1
SESSION LIST
NAME: CP31742 DOMAIN: RAAAN
------------------------------------------------------------------------------
***** PRIMARY ***** **** SECONDARY ****
SEL# NAME TYPE DOM NAME TYPE DOM START TIME END TIME
( 1) RAA CP RAAAN CP31742 CP RAAAN 11/16 18:16:53 *** ACTIVE ***
( 2) CP31742 CP RAAAN RAA CP RAAAN 11/16 18:16:52 *** ACTIVE ***
Figure 35. NLDM List of Sessions between RAA and CP31742
Figure 36. NLDM Display of CP Session 1
APPNCOS N/A
SUBACOS N/A
LOGMODE CPSVCMG 4
PADJ CP CP31742
SADJ CP N/A
Figure 37. NLDM Display of CP Session 2
Note that the mode and APPN COS used here are both CPSVCMG 4. This is
required for, and unique to, CP-CP sessions. The APPN COS for the CONLOSER
session is not relevant to RAA because the 3174 chose it and calculated the
route.
To configure CM/2 as a network node, select SNA local node characteristics from
the configuration list as shown in Figure 40 on page 77.
Select the Network Node option from the next window as shown in Figure 41,
and fill in the fully qualified CP name (NetID and node name).
Next, go back and select SNA connections from the configuration list shown in
Figure 40 to define a logical link to the host. In both CM/2 and CS/2, there are
Either will do the job; we used the first method throughout our testing.
From the connection list shown in Figure 42, create a new token-ring connection
(or edit an existing one), with the partner node type being a host.
The option for APPN support has to be selected, otherwise no CP-CP sessions
will be established. The local PU name has only local significance, and does not
have to match VTAM′s PU name.
Session Setup: The VTAM switched major node (and Communications Manager,
if necessary) are recycled to effect the change. A display of the topology for all
NNs now shows CP05147 with CP-CP sessions to RAA 2, as in Figure 44.
* RAAAN D NET,TOPO,LIST=NN
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.CP31744 NN 128 NONE YES *NA*
IST1296I USIBMRA.CP31743 NN 128 NONE NO *NA*
IST1296I USIBMRA.CP05160 NN 128 NONE NO *NA*
IST1296I USIBMRA.CP31742 NN 128 NONE NO *NA*
IST1296I USIBMRA.CP31741 NN 128 NONE NO *NA*
IST1296I USIBMRA.CP05147 NN 128 NONE 2 YES *NA*
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST1296I USIBMRA.CP31745 NN 128 NONE NO *NA*
IST314I END
Figure 44. Display of Topology for NNs
We test the APPN route to RAA using APING, an LU 6.2 application on the CM/2
workstation. The display of CP05147 shows two additional parallel sessions (as
well as the CP-CP sessions) as shown in Figure 45 on page 80 3.
VTAM Definitions: While the default VTAM connection is LEN, the definitions for
CP05150 do not change, as shown in Figure 46 on page 81.
Session Setup: Session setup will be the same as it was in the pure subarea
network. The CM/2 LEN node is not recognized by VTAM as an end node, since
it is not included in the topology display in Figure 49 and no CP-CP sessions are
established (see Figure 50 on page 83).
* RAAAN D NET,TOPO,LIST=EN
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.CP05137 EN *NA* *NA* YES *NA*
IST314I END
VTAM Definitions: Figure 51 shows the definitions for the PU in VTAM. Once
again, we ensure that CONNTYPE is APPN 1 because we are overriding the
start option which defaults such connections to LEN.
VBUILD TYPE=SWNET,
MAXNO=1,
MAXGRP=1
RAATRPU4 PU ADDR=01,
IDBLK=05D,
IDNUM=05137,
CONNTYPE=APPN, 1
MAXOUT=7,
SSCPFM=USSSCS, (V) VTAM
USSTAB=US327X, (V) VTAM
MODETAB=ISTINCLM, (V) VTAM
DLOGMOD=D4C32XX3, (V) VTAM
MAXPATH=1,
PASSLIM=7,
PUTYPE=2
RAATRPH4 PATH DIALNO=0004400001050001,
GRPNM=EG05L01
RAATLU42 LU LOCADDR=2
RAATLU43 LU LOCADDR=3
RAATLU44 LU LOCADDR=4
RAATLU45 LU LOCADDR=5
Session Setup: The VTAM switched major node and (if changed) the
Communications Manager are recycled to effect the changes. A display of the
network topology for adjacent ENs shows that CP05137 has CP-CP sessions with
RAA 1 (Figure 53).
* RAAAN D NET,TOPO,LIST=EN
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAATRCP2 EN *NA* *NA* NO *NA*
IST1296I USIBMRA.CP05137 EN *NA* *NA* 1 YES *NA*
IST314I END
Figure 53. Display of TOPOLOGY with L I S T = E N
CP05137 is an adjacent CP 1, a dynamically defined CDRSC ( 2 and 5), and
an independent LU 8, just as the 3174 NN. Being a CDRSC it does not require
registration 3. The logmode used by default is CPSVCMG again 4, the actual
The NETSRVR value is shown as “not applicable” at 6. This is because, when
CP05137 is viewed as a CDRSC (from the subarea side of VTAM) it is located by
subarea searching and not by APPN searching. On the other hand, when
CP05137 is viewed as a directory entry 12 (from the APPN side of VTAM), it
must be found via its network node server. That NNS is RAA itself 14. The EN
is seen as a “registered EN” 13 in the directory. This means that the EN is
served by this VTAM as an NNS.
CP05137 has the usual two CP-CP sessions with RAA 11. In this particular
display we have also performed an APING which resulted in the other two
sessions shown 10. These are not CP-CP sessions since they have no “/CP”
status modifier.
* RAAAN D NET,ID=CP05137,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = USIBMRA.CP05137 , TYPE = ADJACENT CP 1
IST486I STATUS= ACT/S----Y, DESIRED STATE= ACTIV 2
IST1447I REGISTRATION TYPE = NO 3
IST977I MDLTAB=***NA*** ASLTAB=***NA***
IST1333I ADJLIST = ***NA***
IST861I MODETAB=***NA*** USSTAB=***NA*** LOGTAB=***NA***
IST934I DLOGMOD=CPSVCMG USS LANGTAB=***NA*** 4
IST597I CAPABILITY-PLU ENABLED ,SLU ENABLED ,SESSION LIMIT NONE
IST231I CDRSC MAJOR NODE = ISTCDRDY 5
IST1184I CPNAME = USIBMRA.CP05137 - NETSRVR = ***NA*** 6
IST1044I ALSLIST = ISTAPNPU 7
IST082I DEVTYPE = INDEPENDENT LU / CDRSC 8
IST654I I/O TRACE = OFF, BUFFER TRACE = OFF
IST1500I STATE TRACE = OFF
IST171I ACTIVE SESSIONS = 0000000006, SESSION REQUESTS = 0000000000
IST206I SESSIONS:
IST1081I ADJACENT LINK STATION = RAATRPU4 9
IST634I NAME STATUS SID SEND RECV VR TP NETID
IST635I RAA ACTIV-S ECC35935B478D526 0004 0003 0 0 USIBMRA 10
IST635I RAA ACTIV-S ECC35935B378D526 0001 0001 0 0 USIBMRA
IST635I RAA ACTIV/CP-S ECC35935B078D526 0007 0001 0 0 USIBMRA 11
IST635I RAA ACTIV/CP-P F7FF6164D4861C5B 0001 0007 0 0 USIBMRA
IST924I -------------------------------------------------------------
IST075I NAME = USIBMRA.CP05137 , TYPE = DIRECTORY ENTRY 12
IST1186I DIRECTORY ENTRY = REGISTERED EN 13
IST1184I CPNAME = USIBMRA.CP05137 - NETSRVR = USIBMRA.RAA 14
IST314I END
Figure 54. Display of EN CP05137
We define the node as a network node and provide the necessary information as
seen in Figure 58 on page 88.
We then select SNA connections as before to get Figure 59, which allows us to
define a new link or change the existing logical link definition to the host.
We are then presented with the adapter list (Figure 60 on page 89) where we
select the required adapter and click on Continue.
The next panel is the Connection to a Host window (Figure 61); clicking on
Additional parameters yields the window shown in Figure 62 on page 90 where
we can select APPN support.
To configure a gateway, select Gateway Hosts and Host LU Pools from the
Configuration List window. The details of gateway configuration are beyond the
scope of a book on APPN; see the CS/2 or CM/2 documentation for advice on
how to do this.
Session Setup: The NCP and the Communications Server machine are recycled
to implement the changes. A display of both the link station (Figure 63) and the
control point LU (Figure 64 on page 91) show that CP-CP sessions have been
established 6 between CP05160 and RAA.
* RAAAN D NET,ID=RA8CM2A,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = RA8CM2A , TYPE = PU_T2.1
IST486I STATUS= ACTIV--L--, DESIRED STATE= ACTIV
IST1043I CP NAME = CP05160 , CP NETID = USIBMRA , DYNAMIC LU = NO
IST1105I RESOURCE STATUS TGN CP-CP TG CHARACTERISTICS
IST1106I RA8CM2A AC/R 21 YES 982D0000000000000000017100808080
IST1482I HPR = NO - OVERRIDE = N/A - CONNECTION = NO
IST081I LINE NAME = LCM2A , LINE GROUP = G08XLLL , MAJNOD = RA8NCPX
IST654I I/O TRACE = OFF, BUFFER TRACE = OFF
IST1500I STATE TRACE = OFF
IST355I LOGICAL UNITS:
IST080I RA8CM2A2 ACTIV RA8CM2A3 ACTIV RA8CM2A4 ACTIV
IST080I RA8CM2A5 ACTIV
IST080I CP05160 ACT/S----Y 6
IST314I END
Figure 63. Display of PU RA8CM2A
Note that each adjacent CP with which VTAM sets up sessions is added
dynamically to the major node ISTCDRDY. Figure 65 shows four of them: the
three PCs and the 3174. The LEN CM/2 node is not represented. Its CP LU has
been predefined and therefore lives in another major node.
* RAAAN D NET,ID=ISTCDRDY,E
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST075I NAME = ISTCDRDY , TYPE = CDRSC SEGMENT
IST486I STATUS= ACTIV , DESIRED STATE= ACTIV
IST478I CDRSCS:
IST483I RAKTX082 ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I CP05160 ACT/S----Y, CDRM = ***NA***, NETID = USIBMRA
IST483I RAKTX093 ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I RAKTX092 ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I RAKAN ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I RAKTX090 ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I SCHANLUC ACT/S----Y, CDRM = RAK , NETID = USIBMSC
IST483I RAKANLUC ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I RAIANLUC ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I RABANLUC ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I RAKTX066 ACT/S----Y, CDRM = RAK , NETID = USIBMRA
IST483I CP05147 ACT/S----Y, CDRM = ***NA***, NETID = USIBMRA
IST483I CP05137 ACT/S----Y, CDRM = ***NA***, NETID = USIBMRA
IST483I CP31742 ACT/S----Y, CDRM = ***NA***, NETID = USIBMRA
IST1500I STATE TRACE = OFF
IST314I END
Figure 65. Display of ISTCDRDY
VTAM Definitions: In this scenario, VTAM is not aware of the CS/2 node, except
as some remote LUs. The CP (CP05221) can be reached using normal APPN
mechanisms, while the dependent LUs appear to be on the link station
RA8CM2A which also connects to CP05160. These dependent LUs are defined
under the PU RA8CM2A, as in Figure 56 on page 87.
From the Configuration List panel, we select Local Node Characteristics and
define the node type, the CP name and the IDBLK/IDNUM (which are not
necessary as VTAM will never see them).
This configuration path will result in CP05221 choosing CP05160 as its network
node server, because it is the only link available to an APPN network node. If
there are multiple NN links on an EN, we can influence the choice of NN server
but we must take the other configuration path:
• SNA Connections to reach the Connection List
• To Network Node to get the list of NN connections
• The connection we are interested in, to get the Adapter List
• Continue to get the Connections to Network Node panel where we enter the
destination MAC address
• Additional Parameters where we select Use this network node connection as
your preferred server and (if the NNS is also VTAM or a gateway, as now)
Solicit SSCP-PU session
Session Setup: When we restart CP05221 the connection is made at once and
CP-CP sessions are established between CP05221 and CP05160. We can now
set up two very different kinds of session between RAA and CP05221:
• Independent LU 6.2 sessions (for example, using APING) take an APPN path
between the two control points RAA and CP05221, using CP05160 as an
intermediate routing node. VTAM RAA sees CP05221 as an ILU/CDRSC, on
link station RA8CM2A, but not an adjacent CP.
• Dependent LU sessions (for example, logging on to TSO) take the same
physical path, but using CP05160 as an interchange between VTAM and
CP05221. Thus VTAM sees the dependent LUs as being owned by itself and
At this stage, when all the resources have been converted to APPN, you can
convert the VTAM start option CONNTYPE to APPN. Subsequently you can let
any new end nodes be connected using native APPN. If a new node attached to
the APPN network can only be configured as a network node, you should check
whether it has enough power and storage to participate in what could be a large
APPN network. The 3174, in particular, will often have to be isolated from the
backbone network because of its limited resources.
At the time of migrating the CMC host to an interchange node, the data host is
not touched at all; that is, it remains as a pure subarea node. All connections to
and from the data host are still subarea connections. Once the interchange
node migration has been completed, the next step is to migrate the data hosts.
For a discussion on the roles of VTAM hosts in an APPN network, refer to 3.1.2,
“VTAM/NCP and APPN Node Roles” on page 26.
In our test network, in addition to the CMC or interchange node RAA, there is a
VTAM data host RAS. In this chapter we discuss the migration of this VTAM
data host to a migration data host (MDH). We then look at converting the
migration data host to a pure end node and discuss why a user might want to do
that. Even though we only use one VTAM data host in the test network for
migration to an MDH, the migration process will be the same for a real network
with numerous VTAM data hosts.
Note: If you only have one VTAM host, then the discussion in this chapter does
not have relevance to your network. A single VTAM should always be converted
to an NN or an ICN. If it becomes an EN or an MDH, it will not be able to take a
full part in your APPN network. It will be unable to own an NCP, and it will not
be able to establish dependent LU sessions unless the dependent LUs are
directly attached to it (because it needs session services extensions in its NN
server).
In order to convert the VTAM data host to an APPN-capable migration data host,
certain VTAM start options and major node definitions have to be modified or
added. For a discussion on what must be considered while planning the data
host migration, refer to Chapter 3, “Planning for Migration” on page 25.
CONNTYPE=LEN 1
CPCP=YES 2
DYNADJCP=YES 3 DEFAULT
HPR=NONE 4
HOSTSA=28 5 << no change
NODETYPE=EN 6
SORDER=SUBAREA 7
VRTG=NO 8 DEFAULT
3 We will allow adjacent control point (ADJCP) minor nodes to be created
dynamically and placed in the major node ISTADJCP.
4 At this point we do not want this VTAM to provide any HPR support.
5 For a migration data host we need to keep this start option and also add
NODETYPE 6.
7 We want to search the subarea network first when a search request
originates on the MDH. In this particular example we wish to ensure that the
proven methods work before making changes to the search patterns.
An MDH will never receive a request from another subarea node and forward it
into the APPN network.
8 In the start options we specify VRTG=NO, and override this value by coding
the VRTG operand for specific CDRM definition statements. By doing this we
can control which SSCP-SSCP sessions will support VR-based transmission
group connections. When VRTG is specified on the CDRM statement, it is also
easier to identify whether an SSCP-SSCP session is supporting VR-TG
connections or not.
For full details on all the APPN-related VTAM start options available, refer to
Appendix B, “VTAM APPN and APPN-Related Start Options” on page 197.
In our network we change the definitions on VTAMs RAA and RAS. On each of
the two VTAMs:
• Change the CA major node for MPC to a local SNA major node with a TRLE
operand pointing at the transport resource list element (TRLE) name.
• Create a TRL major node with LNCTL=MPC defined. This contains the
TRLEs for your MPC connections. The transport resource list element is not
a resource, but describes the connectivity characteristics of the multipath
channel that is being used by the ANNC.
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAS (SA28) TO RAA (SA10) *
****************************************************************
RASAHHC VBUILD TYPE=TRL
*
RASAHHCT 1 TRLE LNCTL=MPC,
READ=(C14), 3
WRITE=(C12) 4
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAA (SA10) TO RAS (SA28) *
****************************************************************
RAALCL2 VBUILD TYPE=LOCAL
RAAPUAHC PU TRLE=RAAAHHCT 1
CONNTYPE=APPN,
CPCP=YES,
TGP=CHANNEL, 2
ISTATUS=ACTIVE
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAA (SA10) TO RAS (SA28) *
****************************************************************
RAAAHHC VBUILD TYPE=TRL
*
RAAAHHCT 1 TRLE LNCTL=MPC,
READ=(C12), 4
WRITE=(C14) 3
Notes:
1 The value on the TRLE= operand must match the label on the TRLE
definition.
3 The READ address C14 in RAS must be connected to the WRITE
address on the other side (that is, in RAA).
4 The WRITE address C12 in RAS must be connected to the READ
address on the other side (that is, in RAA).
For our migration data host RAS, we set up the network node server list major
node as in Figure 72.
*******************************************************************
* NETWORK NODE SERVER LIST MAJORNODE FOR RAS *
* USES RAA AS ITS NETWORK NODE SERVER *
*******************************************************************
RASNNSVR VBUILD TYPE=NETSRVR, X
ORDER=FIRST 1 DEFAULT
RAA 2 NETSRVR NETID=USIBMRA, X
SLUINIT=REQ DEFAULT
3 NETSRVR SLUINIT=REQ DEFAULT
Notes:
1 This is the default and specifies that the EN always attempts to find an
NNS from the NNS list starting with the first entry.
2 We prefer our ICN RAA to be the NNS for this EN.
3 The nameless entry acts as a backup, in case none of the listed
servers are active. Any other adjacent NN, not included in the NNS list,
will be considered as a potential NNS for this EN. However, we ensure
that any NN chosen as a server will support session services extensions,
by coding SLUINIT=REQ. Without session services extensions no
dependent LU sessions can be established between the VTAM EN and the
rest of the network.
On a VTAM APPN host, the default for applications (APPL minor nodes) is to
register at the central directory server, and the default for dependent LUs is to
Resources that will never be search targets should not be registered as this
wastes directory space and network setup time. Such resources include TSO
user applications and NetView operator tasks. By default TSO user applications
are not registered, but you may wish to change the default
(REGISTER=CDSERVR) for other such applications.
The resources on our migration data host, RAS, are all either applications or
dependent LUs; for these resources we do not need to specify the REGISTER
operand on the definition statements because the defaults are acceptable. In
fact, in our network the NNS is the CDS so there is only one registration step for
each resource owned by RAS. RAS does not have any LEN ILUs defined, but if
we create one, it will have to be registered with RAA if it is to be the target of an
APPN search.
Figure 74 on page 101 shows that there are two subarea paths available
between RAA and RAS. One is directly to RAA (subarea 10 1) and the other is
When the CP-CP sessions between ICN VTAM and the MDH VTAM are
established successfully, we get the following messages on RAS:
The display of the TRLE defined for this ANNC link, Figure 78 on page 103,
shows that the READ 1 and WRITE 2 devices defined for the ANNC link are
both active and online. This display also shows the name of the link station 3.
RAS as a migration data host establishes both CP-CP and SSCP-SSCP sessions
with RAA. When we display RAS from itself, Figure 79 shows that is is both a
CDRM 4 and a control point 5. It has both an SSCP-SSCP session 6 and a
pair of CP-CP sessions 7 8 with its partner.
The CP-CP sessions show up as one contention winner 7 (from RAS′s point of
view) and one contention loser 8.
* RASAN D NET,ROUTE,DESTSUB=10
RASAN IST097I DISPLAY ACCEPTED
′ RASAN
IST535I ROUTE DISPLAY 1 FROM SA 28 TO SA 10
IST808I ORIGIN PU = ISTPUS28 DEST PU = ***NA*** NETID = USIBMRA
IST536I VR TP STATUS ER ADJSUB TGN STATUS CUR MIN MAX
IST537I 0 0 INACT 0 1 10 1 INOP
IST537I 0 1 INACT 0 10 1 INOP
IST537I 0 2 INACT 0 10 1 INOP
IST537I 1 0 INACT 1 2 8 1 ACTIV3
IST537I 1 1 INACT 1 8 1 ACTIV3
IST537I 1 2 ACTIV 1 8 1 ACTIV3 31 30 90
IST537I 2 0 INACT 2 2 8 1 ACTIV3
IST537I 2 1 INACT 2 8 1 ACTIV3
IST537I 2 2 INACT 2 8 1 ACTIV3
IST537I 3 0 INACT 3 1 10 1 INOP
IST537I 3 1 INACT 3 10 1 INOP
IST537I 3 2 INACT 3 10 1 INOP
Figure 80. Subarea Paths with RAS as an MDH
All the direct subarea paths (those with subarea 10 adjacent) are now INOP 1,
while those through subarea 8 are carrying the SSCP session 2. Compare this
with Figure 74 on page 101.
The display of the primary LU RASAT01 as seen from the ICN RAA is shown in
Figure 83 on page 106.
The display provides us with a lot of new information about the session, which is
now using an APPN connection.
1 shows that RASAT01 is a CDRSC, just as it would have been in the subarea
environment. However, its owning VTAM is shown as RAA 2 rather than RAS.
This implies that RASAT01 is seen by RAA as an independent LU, which fact is
confirmed in the message IST082I 4. The adjacent link station list is ISTAPNPU
3, meaning that RAA expects to find this independent LU using the APPN
network. The actual link station being used is RAAPUAHC 5, which is the
ANNC link between the two VTAMs.
The perceptive reader will ask why the session has taken an APPN path when
SORDER=SUBAREA was specified in the start options on both VTAMs. The
reason is that RASAT was registered to RAA by RAS when the CP-CP sessions
were started. The original session request came from RAA′s subarea domain,
and RAA searched its local directories before trying any network searching. Its
subarea directory had no knowledge of RASAT, but its APPN directory did. Thus
RAA knew RASAT as an APPN resource and sent an APPN Locate to RAS to
verify its availability.
The NetView Session Monitor display for the same session, Figure 84, provides
us with APPN-related information on the session characteristics:
• The primary LU for the session is reached via the link station (RAAPUAHC)
3 representing the ANNC connection, as seen in the previous display.
• The mode entry M2SDLCQ has neither subarea COS nor APPN COS defined.
Thus there is no subarea COS defined for the session 4, but VTAM has
used the default APPN COS #CONNECT 5 for the APPN part of the session.
A blank (or missing) COS is not permitted in APPN. For a further discussion
of how VTAM selects a class of service when a session traverses both APPN
and subarea networks, please see Chapter 10, “Logmode and COS
Resolution in a Mixed Network” on page 167.
• The adjacent CP at the PLU side of the session is RAS 6 (an APPN node)
and the adjacent CP at the SLU side is CP05137 7. Although CP05137 has
an APPN connection to RAA, this session does not use it and it does not
appear in the NetView APPN Route display (Figure 85 on page 108). The
dependent LU RAATLU44 is directly attached to the NCP boundary function
and uses peripheral subarea protocols for the route extension to CP05137.
The Session Monitor APPN route (AR) display, Figure 85 on page 108, shows the
APPN route taken by the session. 8 is APPN TG21 between the VTAMs, the
ANNC connection.
If you wish to code a preferred TG number, then both CPNAME= and NETID=
parameters must also be coded on the link station definitions.
+---------+
| CP |
|RAS |
+----+----+
TG021 | 8
+----+----+
| CP |
|RAA |
+---------+
Unless there is a specific need to retain a data host as an MDH, it can be made
a pure end node. For example, if VTAM must support BSC 3270 sessions then it
requires a subarea path into the network and must therefore remain an MDH.
Note
The one thing you should never do is have independent subarea and APPN
paths between the same VTAM nodes. This leads to unpredictable session
routes and possibly session failures. If you require both subarea and APPN
connections between two nodes, use a VR-TG to integrate the two.
The new local SNA definitions created on VTAM RAS for the NCP type 2.1
connection are shown in Figure 86. XID=YES is required for a type 2.1
connection.
Note: A reader of the previous edition of this book has kindly pointed out that
the definitions should include a MAXBFRU keyword. MAXBFRU now defaults to
5 but used to default to 1 until very recently. Unless your I/O buffers are very
large, a MAXBFRU of 1 is likely to result in a connection failure as soon as a
reasonably large PIU arrives. MAXBFRU=15 is more appropriate for an NCP
connection.
*******************************************************************
* CA MAJNODE FOR CHANNEL CONNECTION (TYPE2.1) TO RA8NCPX *
* (FID2 connection) *
*******************************************************************
RASCA8 VBUILD TYPE=LOCAL
RASE20P PU PUTYPE=2,
CUADDR=E20,
XID=YES,
CPCP=YES,
CONNTYPE=APPN
The definition changes required in the NCP RA8NCPX to convert the NCP
channel link to a FID2 connection are shown in Figure 87 on page 110. The
keyword PUTYPE is changed from PUTYPE=5 to PUTYPE=2 1 and the
keyword XID=YES 2 has been added.
Note that ISTATUS=INACTIVE 3 on the GROUP definition is not desirable for a
type 2.1 connection. For a subarea connection to a host INACTIVE is required,
***********************************************************************
******** LNCTL=CA definitions for NCP RA8NCPX **********************
***********************************************************************
RA8GCA GROUP LNCTL=CA, X
ISTATUS=INACTIVE3 STOP VTAM ACTIVATING CHANNEL LINK
RA8CE20 LINE ADDRESS=0, 1ST CA. PHYSICAL POSTION 5 X
CA=TYPE6-TPS, 3745 CHANNEL ADAPTOR TYPE X
CASDL=120, INTERVAL BEFORE CHANNEL SLOWDOWN X
CONNTYPE=APPN, APPN CONNECTION X
DELAY=0.2, CHAN ATTN DELAY X
DYNADMP=NONE, NO EP SUBHANNELS TO DUMP DATA OVER X
NCPCA=ACTIVE, NATIVE SUBCHANNEL(NSC) ACTIVE X
TIMEOUT=120 INTERVAL BEFORE CHANNEL DISCONTACT
RA8PCA PU PUTYPE=2, 1 PERIPHERAL TYPE 2.1 CONNECTION X
XID=YES, 2 FOR TYPE 2.1 FID2 X
TGP=CHANNEL, TG profile X
TGN=1
When we display RAS from itself (see Figure 89 on page 111), we see that it is a
CP only 3, and not a CDRM. It has CP-CP sessions with RAA but no SSCP
session. A display of RAS from RAA would show RAS as an adjacent CP and
not a CDRM.
There are no subarea paths to display as RAS no longer has the ability to
establish ERs and VRs. Although RAS has no “externally known” subarea
number, it still has an internal number. All VTAM pure NNs and ENs use a
number of one (1) for the purpose of network address assignment. VTAM traces
will continue to show network addresses with subarea number 1 being used. All
remote resources (CDRSCs) will have element addresses associated with
subarea 1.
With RAS as an end node, there are two APPN FID2 connections between the
ICN RAA and the end node RAS, as shown in Figure 90 on page 112. The direct
channel link from RAS to RAA is TG 21, and the channel link from RAS to
RA8NCPX is TG 22. The TG numbering simply reflects the order in which the
two links were activated.
The tests conducted for LU-LU sessions showed that the same sessions could be
established with data host RAS as an end node, as were possible in the
traditional subarea network and RAS as an MDH.
The display of TRLE RASAHHCT on RAS (Figure 92 on page 114) shows that the
I/O READ and WRITE devices are still 2 offline and in RESET status. When the
ANNC link between RAA and RAS is successfully activated, the TRLE display is
as in Figure 78 on page 103, with the READ/WRITE devices showing as active
and online.
The other mistake we made was not to point the local SNA PU to the correct
TRLE, as discussed in 5.2, “Converting Subarea MPC to ANNC” on page 96.
This causes an activation error with the sense code 08170009.
If your 3745(s) are heavily utilized, there are two things you could do:
• Migrate direct to HPR from subarea. HPR uses minimal resources in
intermediate routing nodes (which an NCP always will be for an HPR
connection) because there is no session awareness in such nodes.
• Implement VR-TG over the subarea connection. This is transparent to the
NCP, but means you will still have to maintain the appropriate subarea path
tables.
The APPN border node function is roughly equivalent to subarea SNI, but the
differences are significant. For instance, an SNI border is inside an NCP
whereas an APPN border is between two nodes. The functions and benefits of
an APPN border node are:
• The border node function allows two network nodes with different NetIDs to
connect to each other. This connection provides directory services for the
search process and LU-LU connectivity.
• The border node function allows the partition of a single network into
subnetworks (the subnetworks having the same NetID).
• The border node function allows a network with a single NetID to be
partitioned into several parts, separated by networks of differing NetIDs.
• The border node function allows two network nodes, with the same or
different NetIDs, to isolate their network topologies from one another. This
results in the reduction of network overhead and allows nodes with limited
resources to participate in APPN networking. It also reduces the storage
requirement for the topology database in each of the network nodes within
the subnetwork.
• The APPN COS definitions on both sides are isolated. This reduces the
administrative requirement for coordinated definitions.
• The extended border node function between two VTAM network nodes is
activated with minimal definitions. The border nodes use dynamic processes
to exchange their capabilities and information on reachable networks.
• Cross network connection can be made through channel-to-channel links, as
well as any other APPN connection where at least one partner supports the
border node function. There is no need for an NCP, although NCP
connections can of course be used.
• Sessions can be established across the subnetwork boundaries. Searches
can cross borders but since the network topologies are independent the
route chosen for a session across a border may not be the optimum route.
There are two types of border node in APPN, and two types of border . The node
types are:
• An extended border node (EBN) provides for the multi-subnetwork
environment the full function available within a single network. Some of the
functions may not work as efficiently as within one network, but there are no
restrictions on network design, and any or all session paths can be used.
Today only VTAM (as NN, ICN or CNN) can be an extended border node,
although 2216 support has been announced.
• A peripheral border node (PBN) provides a limited set of functions. For
example:
− A PBN does not support session services extensions.
− The DLUR/DLUS pipe cannot cross a border managed by a PBN, nor can
any session from a DLUR dependent LU.
− Therefore, a peripheral border node supports only PLU-initiated LU 6.2
sessions across a border it manages.
− A PBN cannot partition a network into multiple pieces with the same
NetID.
For our test network, we make changes to the network definitions to create
APPN subnetworks with border node connections, as in Figure 93 on page 118.
We are connecting an EBN (VTAM) to various network nodes that do not support
border node functions, therefore we are creating peripheral borders managed by
an extended border node . These borders have restrictions on dependent LU
traffic, because not all of the partner NNs support session services extensions.
Also, they must be at the ends of any session path that crosses them. However,
these factors are not of concern in our environment because:
• Our dependent LUs are still directly connected to VTAM boundary functions,
and therefore obey the rules. Their sessions do not cross any borders.
• All our borders are managed by the same VTAM, so they will always be at
the ends of a session path.
You should avoid the use of peripheral borders when connecting two VTAMs (by
configuring one VTAM as an EBN and the other as an NN). Dependent LU
sessions will work (because both partners support session services extensions),
but the other restrictions on peripheral borders still apply. Thus:
• A peripheral border cannot be an intermediate border on a session path that
traverses more than three subnetworks.
• Multiple peripheral borders between the same subnetworks will not work
reliably.
• If you use DLUR/DLUS, then:
− If the DLUR/DLUS sessions cross the border, the EBN must be on the
DLUS side of the border.
− The dependent LU′s session partner (the PLU) must be connected to the
DLUS by an extended border (two EBNs).
Figure 93 on page 118 illustrates how we set up the APPN borders in our test
environment.
Note:
• If you use an NCP to provide the boundary function for the intersubnetwork
connection (in other words, if one side of the APPN border is an NCP) you
need NCP V7R1 or later.
• If you use HPR across an APPN border where one side (or both) is an NCP,
you need NCP V7R5 or later.
• An APPN border cannot be set up across a connection network. In other
words, any connection between two network nodes with different NetIDs
must be explicitly defined.
Notes:
1 is required to specify that this VTAM node is to provide the extended
border node function.
2, 3, and 4 are meaningful only when BN=YES is specified.
2 specifies that all active border nodes, native or nonnative, may be
added to the routing list (adjacent cluster table), provided certain
conditions are met. These conditions effectively allow dynamic rerouting
of search requests only when the target LU has the NetID of an adjacent
network, or when a previous incoming search has identified where that
NetID is located. BNDYN is roughly equivalent to SSCPDYN in an SNI
environment.
4 specifies that this border node VTAM will search a maximum of three
subnetworks, when looking for a resource. To allow a search request to
traverse at least one APPN border, a value of SNVC=2 is required.
SNVC=1 restricts a search to just its native network.
For further details on these VTAM start options, refer to the chapter “Start
Options” in the VTAM Resource Definition , SC31-8565.
For our network, we set up an adjacent cluster routing list for border node RAA,
as in Figure 95 on page 120.
Figure 95. Adjacent Cluster Routing List for RAA Border Node
Notes:
5 is the routing list (with no NETID parameter), which will be used as a
default when:
• A non-network qualified request is received
• A network qualified request is received and the network identifier is
not defined in any other NETWORK statements
6 is the routing list used when USIBMRA is the network ID specified on
the request.
7 is the routing list used when USIBMRAX is the network ID specified on
the request.
Coding BNDYN=NONE will ensure that VTAM does not forward any search
unless the target network is predefined in the adjacent cluster table.
Thus, looking at the entry 6 for USIBMRA (the native network):
• If the 3174 (CP31742) sends in a search for USIBMRA. node , RAA will search
its own network. It will not forward the search to CP31742 for obvious
reasons. Nor will it forward the search to the CS/2 node (CP05160), despite
Without any extra definitions, VTAM will establish a peripheral border between
itself and CP05160, simply because it will discover a non-native NetID on the
adjacent connection.
Figure 96. NCP Definition Change for Network Node CP31742 (3174)
RA8CM2A PU ADDR=C1,
...
PUTYPE=2,
NETID=USIBMRAX,8 FOR SUBNETWORK CONNECTION
XID=YES, FOR T2.1 NODE
...
Figure 97. NCP Definition Change for Network Node CP05160 (CS/2)
VTAM displays confirm that RAA is now defined as a border node 1 (refer to
Figure 98 on page 123) but as it has no cross-border connections, it is not acting
as one yet ( 2 and 3 in the topology display in Figure 99 on page 123).
* RAAAN D NET,TOPO
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1306I LAST CHECKPOINT ADJ NN EN SERVED EN CDSERVR ICN BN
IST1307I 11/28/95 17:09:35 1 2 4 2 1 1 0 2
IST314I END
* RAAAN D NET,TOPO,ID=RAA,LIST=ALL
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST1297I ICN/MDH CDSERVR RSN
IST1298I YES YES 202
IST1223I BN NATIVE
IST1224I NO 3 YES
Figure 99. Network Topology without any Border Node Connections
With the ADJCLUST definitions (see Figure 95 on page 120) now active, we
display the adjacent cluster list on RAA. Refer to Figure 100 on page 124 for
details. Note that the entries defined in the adjacent cluster list (Figure 95 on
page 120) are listed with a TYPE of DEFINED 4; that is, they are the predefined
entries in the adjacent cluster table. Because the links to the two nodes,
CP31742 and CP05160, have not yet been activated from RAA, they both are
shown as NOT ACTIVE 5.
Next, we activate the connections from RAA to the two nodes, CP31742 and
CP05160, and establish CP-CP sessions between RAA and each of the two
nodes.
Once the CP-CP sessions are active, we display the network topology on RAA
and find that some changes have occurred. Refer to Figure 101 on page 125 for
details. The topology display now shows that there is one border node 6
active in this subnetwork. We can also see that this border node is RAA itself
7.
From the list of APPN transmission groups originating at RAA, we can also see
that the links to CP31742 and CP05160 now have a TGTYPE of INTERCLUST 8;
that is, they are links to different APPN clusters or subnetworks.
* RAAAN D NET,TOPO,ID=RAA,LIST=ALL
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST1297I ICN/MDH CDSERVR RSN HPR
IST1298I YES YES 212 NONE
IST1223I BN NATIVE
IST1224I YES7 YES
IST1299I TRANSMISSION GROUPS ORIGINATING AT CP USIBMRA.RAA
IST1357I CPCP
IST1300I DESTINATION CP TGN STATUS TGTYPE VALUE WEIGHT
IST1301I USIBMRA.RAS 21 OPER ENDPT YES *NA*
IST1301I USIBMRA.RAS 255 OPER ENDPT VRTG YES *NA*
IST1301I USIBMRA.CP31742 21 OPER 8 INTERCLUST YES *NA*
IST1301I USIBMRA.CP05147 21 OPER INTERM YES *NA*
IST1301I USIBMRA.CP05150 21 OPER ENDPT NO *NA*
IST1301I USIBMRA.CP05137 21 OPER ENDPT YES *NA*
IST1301I USIBMRAX.CP05160 21 OPER 8 INTERCLUST YES *NA*
IST314I END
Figure 101. Topology Display with Border Node Connections
The display for the adjacent cluster list (refer to Figure 102 on page 126) now
shows both the nodes, CP31742 and CP05160, as ACTIVE 1. The STATUS field
in the display shows the result of the last search sent to this node. Because no
search requests for a resource have been sent to these nodes, the status is
correctly reflected as NOT SEARCHED 2.
Another point worth noting is the fact that the two CPs have been dynamically
added 3 to the default adjacent cluster table, and can be used for future
session routing. The dynamic addition of CPs to the adjacent cluster table is
dependent on the value of the BNDYN parameter in the VTAM start options or in
the ADJCLUST table. Since we have BNDYN=LIMITED VTAM has added the
entries for the adjacent CPs to the default table. The default table is used for
non-network-qualified requests, so all adjacent NetIDs are potentially valid
search targets for such requests.
With the links now active between subnetworks, we test the role of the adjacent
cluster list while carrying out search requests. We make use of the VTAM D
NET,APING,ID= command to initiate a search request for a remote resource. Due
to the manner in which VTAM RAA start options have been defined, it will search
both the subarea and APPN networks for an unknown resource. In this section,
we are only interested in the APPN search.
Because we have specified a fully qualified resource name, VTAM will use only
the adjacent cluster table for NetID USIBMRAX. Once the session setup
succeeds, we display the adjacent cluster table (refer to Figure 104).
* RAAAN D NET,ADJCLUST,SCOPE=ALL
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = ADJACENT CLUSTER TABLE
IST1325I DEFINED TABLE FOR DEFAULT_NETID DYNAMICS = LIMITED
IST1326I CP NAME TYPE STATE STATUS SNVC
IST1327I USIBMRA.RAA DEFINED ACTIVE *** N/A *** N/A
IST1327I USIBMRA.CP31742 DYNAMIC ACTIVE NOT SEARCHED 003
IST1327I USIBMRAX.CP05160 DYNAMIC ACTIVE NOT SEARCHED 003
IST924I --------------------------------------------------------
IST1325I DEFINED TABLE FOR USIBMRA DYNAMICS = LIMITED
IST1326I CP NAME TYPE STATE STATUS SNVC
IST1327I USIBMRA.RAA DEFINED ACTIVE *** N/A *** N/A
IST1327I USIBMRA.CP31742 DEFINED ACTIVE NOT FOUND 003
IST924I --------------------------------------------------------
IST1325I DEFINED TABLE FOR USIBMRAX DYNAMICS = LIMITED
IST1326I CP NAME TYPE STATE STATUS SNVC
IST1327I USIBMRAX.CP05160 DEFINED ACTIVE FOUND 5 003
IST314I END
Figure 104. ADJCLUST Table Status after Test 2
The display shows a change in the STATUS field for node CP05160, from NOT
SEARCHED to FOUND 5. This implies that the node CP05160 has been
searched and the resource was found.
Use the product-supplied APPN COS names as much as possible. Often the
APPNCOS=#CONNECT start option will serve you well.
If you have a large network and require a tighter control over cross-network
searching, then define an adjacent cluster table and use BNDYN=LIMITED or
even BNDYN=NONE. BNDYN=FULL is recommended only for small networks
where a search sent out to all adjacent subnets will have minimal impact.
Define the SNVC (subnet visit count) to an appropriate value, such that it covers
all the search paths you may have in your network. Do not forget to allow for
backup paths in case of a failure; such backup paths may be longer than the
primary paths.
Do not forget that an APPN end node need not have the same NetID as its NN
server. Therefore, converting an SNI-attached data host to an end node removes
a subnetwork boundary without the need to change the NetID.
Do not mix extended and peripheral borders between the same two networks.
This can result in session failures or poor route selection.
If you use peripheral borders to connect a network, have all the connections
owned by a single NN and connect them only to a single EBN in the other
network. This way you will avoid the search looping problems that lead to
session failures and undesirable routes.
The VR-TG actually represents all the possible VRs between the two VTAMs and
their attached NCPs. When an APPN network node calculates a route which
includes a VR-TG between two VTAMs, the VR-TG partner acting as SSCP(PLU)
selects a subarea virtual route using the normal subarea mechanisms. This
selected VR then becomes the actual session path across the VR-TG.
Note
7.1 Definitions
To illustrate VR-TG, we make a small change to the network used in previous
chapters and remove the MPC connection between RAA and RAS. This
represents a typical situation where a remote data host is connected to a
backbone NCP owned by the CMC. If we require subarea access to this host (it
may be a backup CMC, for instance) then the connection between RAS and RAA
must be FID4. If we also require APPN connectivity (for DLUR/S, for example)
then VR-TG is necessary.
In our example we have migrated RAS to an MDH (it cannot be a pure end node
if VR-TG is used) and we will set up a VR-TG connection between RAA and RAS.
Figure 105 depicts the network diagram. The VR-TG between the two nodes
comprises all the VR/TP combinations between:
• Subarea 10 and subarea 28
• Subarea 5 and subarea 28
• Subarea 8 and subarea 28
In our setup, we will use the second option (that is, code VRTG=YES on the
individual CDRM definitions). Note that two VR-TG partners must be adjacent in
the subarea sense; in other words they must have an SSCP-SSCP session. The
SSCP-SSCP session is used to exchange basic APPN connection information that
would flow on XIDs on a FID2 connection.
Figure 106 shows the CDRM definitions we code on the VTAM nodes.
***************************************************
* CDRM MAJOR NODE FOR RAA
* SUBAREA 10 AS AN ICN AND SUBAREA 28 AS AN MDH
***************************************************
VBUILD TYPE=CDRM
NETWORK NETID=USIBMRA
RAA CDRM SUBAREA=10,CDRDYN=YES,CDRSC=OPT
RAK CDRM SUBAREA=20,CDRDYN=YES,CDRSC=OPT
RAS CDRM SUBAREA=28,CDRDYN=YES,CDRSC=OPT,VRTG=YES, 1
VRTGCPCP=YES 2
***************************************************
* CDRM MAJOR NODE FOR RAS
* SUBAREA 10 AS AN ICN AND SUBAREA 28 AS AN MDH
***************************************************
VBUILD TYPE=CDRM
NETWORK NETID=USIBMRA
RAA CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=10,VRTG=YES, 1
VRTGCPCP=YES 2
RAK CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=20
RAS CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=28
Notes:
Each VR-TG is reported to the APPN topology database, and therefore has
TG characteristics just as any other APPN TG. However, VTAM cannot
know the true TG characteristics for a VR-TG and so they default to
CAPACITY=8K and other values which are likely to be undesirable.
2 indicates that CP-CP sessions will be supported over the VR-based
TG. If there are no CP-CP sessions active between the two VTAM nodes,
CP-CP session establishment is automatically initiated when the VR-TG is
activated. A VR-TG is treated the same way as any other APPN
connection: if CP-CP sessions are to be set up then they will use the first
connection to be activated, whether VR-TG or FID2.
* RAAAN D NET,TOPO,ID=RAA,LIST=ALL
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST1297I ICN/MDH CDSERVR RSN HPR
IST1298I YES YES 292 NONE
IST1223I BN NATIVE
IST1224I YES YES
IST1299I TRANSMISSION GROUPS ORIGINATING AT CP USIBMRA.RAA
IST1357I CPCP
IST1300I DESTINATION CP TGN STATUS TGTYPE VALUE WEIGHT
IST1301I USIBMRA.CP05150 21 OPER ENDPT NO *NA*
IST1301I USIBMRA.CP05147 21 OPER INTERM YES *NA*
IST1301I USIBMRA.CP05137 21 OPER ENDPT YES *NA*
IST1301I USIBMRA.CP31742 21 OPER INTERCLUST YES *NA*
IST1301I USIBMRAX.CP05160 21 OPER INTERCLUST YES *NA*
IST1301I USIBMRA.RAS 21 3 INOP ENDPT YES *NA*
IST314I END
Figure 107. Topology Display for RAA with No VR-Based TG
For the test, we use the APING transaction program from a CM/2 network node,
CP05147, to establish LU 6.2 sessions to the VTAM host RAS. The physical
network for the two session partners is as indicated in Figure 108 on page 135.
With no VR-TG activated, we establish the sessions between CP05147 and RAS.
The NetView Session Monitor session configuration display (refer to Figure 109)
shows that the session path goes from CP05147 1 to NCP RA5NCPX 2 and
then via ER 1 3 to its destination LU RAS 4. Figure 110 on page 136 shows
the detailed route for ER 1. This route goes from RA5NCPX (subarea 5) 5 to
RA8NCPX (subarea 8) 6 to RAS (subarea 28) 7. So the complete route taken
is as shown below:
CP05147 - RA5NCPX - RA8NCPX - RAS
Looking at the APPN route configuration display (Figure 111), we get an APPN
view of the session route. We see that the session is going via TG21 from
CP05147 to RAA 8; then from RAA it uses a subarea route to reach its
destination 9. The TG labelled “IN-TG” is an interchange TG. To the APPN
network this (always TG 254) represents a TG to an end node served by the ICN;
in fact it is a subarea route adjacent to an APPN route.
+---------+
| CP |
|CP05147 |
+----+----+
TG021 | 8
+----+----+
| CP(ICN) |
|RAA |
+----+----+
IN-TG | 9
+----+----+
| SUBAREA |
| NODE(S) |
+---------+
Figure 111. APPN Route Configuration with No VR-TG
Next, we turn on the VR-TG by deactivating the respective CDRM major nodes
on RAA and RAS, and reactivating them with the new VR-TG definitions, as
shown in Figure 106 on page 133. The deactivation of CDRM major nodes is
done using the SAVESESS keyword, so that existing LU-LU sessions are not
disrupted.
The APPN topology as seen from RAA (refer to Figure 113) now shows an active
APPN transmission group from RAA to RAS with TGN=255 1 with a TGTYPE of
VRTG. Note that TG21 to RAS is still in the topology database with an INOP
status 2.
* RAAAN D NET,TOPO,ID=RAA,LIST=ALL
RAAAN IST097I DISPLAY ACCEPTED
′ RAAAN
IST350I DISPLAY TYPE = TOPOLOGY
IST1295I CP NAME NODETYPE ROUTERES CONGESTION CP-CP WEIGHT
IST1296I USIBMRA.RAA NN 128 NONE *NA* *NA*
IST1297I ICN/MDH CDSERVR RSN HPR
IST1298I YES YES 292 NONE
IST1223I BN NATIVE
IST1224I YES YES
IST1299I TRANSMISSION GROUPS ORIGINATING AT CP USIBMRA.RAA
IST1357I CPCP
IST1300I DESTINATION CP TGN STATUS TGTYPE VALUE WEIGHT
IST1301I USIBMRA.CP05150 21 OPER ENDPT NO *NA*
IST1301I USIBMRA.CP05147 21 OPER INTERM YES *NA*
IST1301I USIBMRA.CP05137 21 OPER ENDPT YES *NA*
IST1301I USIBMRA.CP31742 21 OPER INTERCLUST YES *NA*
IST1301I USIBMRAX.CP05160 21 OPER INTERCLUST YES *NA*
IST1301I USIBMRA.RAS 21 2 INOP ENDPT YES *NA*
IST1301I USIBMRA.RAS 255 OPER 1 ENDPT VRTG YES *NA*
IST314I END
Figure 113. Topology from RAA with VR-TG
With the VR-TG connection now active, we terminate the session between
CP05147 and RAS, and delete any dynamically created CDRSCs so we start with
a clean directory.
We then reestablish the session between CP05147 and RAS, using the CM/2
APING transaction program.
The NLDM displays for the session configuration and the ER are the same as
shown in Figure 109 on page 135 and Figure 110 on page 136, but the APPN
session route configuration has changed. As you can see from the APPN route
display (Figure 114 on page 138), the end-to-end connection from CP05147 to
RAS now appears as an APPN connection, where VR-TG 3 is being used for
the hop between RAA and RAS.
Note that, due to a bug in VTAM V4R3, RAS is shown as an ICN in the figure. It
is, of course, an MDH.
+---------+
| CP |
|CP05147 |
+----+----+
TG021 |
+----+----+
| CP(ICN) |
|RAA |
+----+----+
VR-TG | 3
+----+----+
| CP(ICN) |
|RAS |
+---------+
Figure 114. APPN Session Route Configuration with VR-Based TG
So for the example discussed, the APPN route will now look like Figure 115.
The net result of setting up VR-based TG between RAA and RAS is that we have
managed to overlay an APPN route over an existing subarea FID4 route between
the two VTAMs. But, more importantly, now the two VTAMs have become part of
the same APPN network and therefore have access to all of the APPN
capabilities such as DLUS and HPR.
In the early stages of migration from subarea to APPN, you may wish to bring up
CP-CP sessions between two VTAMs connected by VR-TG in a controlled
fashion, rather than to allow them to be started at VTAM initialization. You can
certainly change the VRTGCPCP option dynamically, but:
• It requires both partner SSCPs to have the option set, whether by CDRM
activation or by the MODIFY VTAMOPTS command.
• It requires the CDRM major node to be recycled on the VTAM which receives
the SSCP session request.
A better way to achieve control over CP-CP session setup over a VR-TG is as
follows:
• In each VTAM partner, create a CDRSC major node with a definition for the
partner CP, but with ISTATUS=INACTIVE. Add it to ATCCONxx so that it
gets activated early in the startup cycle.
• When the VR-TG is activated the CP-CP sessions will not be established and
only subarea protocols will be used.
For SSCP takeover in a combined subarea and APPN network the takeover
VTAM needs to be defined as an interchange node specifying NODETYPE=NN
and HOSTSA=nn. An MDH, NN or EN cannot participate in takeover processing
because none of them can own an NCP. If you try to take over resources from
an MDH, you will receive the following message:
During the first stages of APPN migration, the takeover VTAM may still be a
subarea-only VTAM. Possible configurations for takeover SSCPs are listed
below and you should take these into account when migrating:
1. The takeover SSCP is still a subarea-only VTAM. All APPN connections
being taken over from the primary SSCP will become LEN connections. No
CP-CP sessions are established until the original SSCP or an APPN-capable
SSCP regains control.
2. The takeover SSCP has implemented APPN and adjacent CPs do not support
CP name change. In this case, if the takeover SSCP is not the original
SSCP, the connections become LEN connections. The connections must be
broken and reestablished (disrupting LU-LU sessions) for the new CP-CP
sessions to start.
3. The takeover SSCP has implemented APPN and adjacent CPs are located
over subnetwork boundaries. The takeover SSCP must define the
connections in the same way as the original SSCP and it must also support
APPN multiple network connectivity (border node). Otherwise the
connections will be LEN connections.
4. The takeover SSCP that has implemented APPN is an ICN and adjacent CPs
support CP name change. This should be your objective.
NCP has the capability of initiating and accepting nonactivation XID3s whether it
assumes the primary or secondary role on a connection. Therefore, you should
ensure that NCP is the primary station. This is already almost always the case
for SDLC links. For NCP-NCP connections (where one NCP must be secondary)
it is of no concern since NCP supports nonactivation XID3 in either role. In fact,
as soon as an NCP has detected the loss of its owner it sends nonactivation
XID3s to adjacent network nodes with the TG quiescing bit turned on. This
informs them that the active link information will change. Once takeover has
been established, the new VTAM sends out topology data updates to the network
to inform it of the changes.
If you are using the Configuration Services XID exit (ISTEXCCS) or the
self-defining dependent LU support exit (ISTEXCSD), copies of these should be
active in the takeover SSCP so that resources dynamically defined using these
exits can be taken over and re-defined.
Note:
6This PS/2 is running CM/2 V1.11 and is a LEN node with a CP name of
CP05150. Of course, no CP-CP sessions exist. The PU and LUs
associated with this node are dynamically defined in RAA using the
ISTEXCCS exit and model major node RAAMDLS. The PU name is
W05150 and the LUs are W0515002-11. W0513702 is a dependent LU in
session with RASAN on RAS.
The difference in this scenario is that we have used dynamic definitions for all
the dependent LU definitions on the PCs and 3174s. That is, we have allowed
the configuration services XID exit ISTEXCCS to define PUs and LUs based on
In our scenario for APPN SSCP takeover, the primary SSCP RAA remains
defined as an ICN as described in Chapter 4, “Migrating a CMC Host to ICN” on
page 61. In order for the backup SSCP to initiate the takeover of network
connections, we must change the definitions on the designated takeover SSCP.
In our case, we have used RAS, which was defined as an MDH and then as an
EN. We have redefined RAS as an ICN for it to be able to perform the takeover.
This node must be able to take over the network and have the same capabilities
as the primary SSCP. This includes defining RAS as a central directory server.
Figure 117 shows the start options we defined for RAS in the takeover
environment.
*******************************************************************
* ATCSTR28 FOR RAS *
* THIS IS SUBAREA 28 AS A BACKUP ICN FOR SSCP TAKEOVER OF RAA *
*******************************************************************
APPNCOS=#CONNECT,
CONFIG=28,
BN=YES,
CDSERVR=YES,
CONNTYPE=APPN,
CONFIG=28,
CPCP=YES,
CSALIMIT=0,
DIRSIZE=0,
DIRTIME=8D,
DYNADJCP=YES,
DYNLU=YES,
GWSSCP=YES,
HOSTPU=ISTPUS28,
HOSTSA=28,
HPR=NONE,
INITDB=ALL,
IOBUF=(100,500,5,,1,30),
MAXSUBA=255,
NETID=USIBMRA,
NODETYPE=NN,
NOTRACE,TYPE=VTAM,IOINT=0,
NUMTREES=100,
RESUSAGE=100,
ROUTERES=128,
PPOLOG=YES,
SORDER=SUBAREA,
SSCPID=10028,
SSCPNAME=RAS,
SSEARCH=YES,
SUPP=NOSUP,
TNSTAT,CNSL,
VRTG=NO,
XNETALS=YES
We were able to view the CP-CP sessions using either the command D
NET,ID=ISTADJCP,E for dynamically defined adjacent CPs or D NET,ID=RAAADJCP,E
for a list which we had statically defined. Figure 118 shows an example.
DISPLAY NET,ID=RAAADJCP,SCOPE=ALL
IST097I DISPLAY ACCEPTED
IST075I NAME = RAAADJCP , TYPE = ADJCP MAJOR NODE
IST486I STATUS= ACTIV , DESIRED STATE= ACTIV
IST1100I ADJACENT CONTROL POINTS FROM MAJOR NODE RAAADJCP
IST1102I NODENAME NODETYPE CONNECTIONS CP CONNECTIONS NATIVE
IST1103I USIBMRA.CP05137 EN 2 1 *NA*
IST1103I USIBMRA.CP05147 NN 1 1 YES
IST1103I USIBMRA.CP05160 NN 2 1 YES
IST1103I USIBMRA.CP31742 NN 2 1 YES
IST1103I USIBMRA.CP31744 NN 1 1 YES
IST314I END
Figure 118. Display of Adjacent CPs
We then issued the V NET,REL VTAM command for both NCPs owned by RAA,
namely RA5NCPX and RA8NCPX. The VARY REL,TYPE=GIVEBACK command
releases the NCP and its resources from VTAM′s knowledge so that another
VTAM can acquire them. Figure 119 on page 147 shows the console log after
the NCPs and their resources were released by RAA.
In fact, apart from the CP-CP session failure messages there is nothing
APPN-specific in these displays. VTAM has simply relinquished ownership of
some dependent LUs and some link stations.
Next, we took over the NCPs from RAS using the commands:
V NET,ACQ,ID=RA8NCPX,ACT,PUSUB
V NET,ACQ,ID=RA5NCPX,ACT,PUSUB
Once again the resulting displays showed nothing APPN-specific except for those
nodes which supported CP name change. In fact there were only three such
nodes in the network, since the CM/2 nodes were not capable of it. Figure 120
on page 149 shows the message we received when the 3174 CP31742
re-established its new CP-CP sessions.
The following diagram is the view of the network after SSCP takeover by RAS.
The connections and the sessions are the same as in the previous network
diagram, except that:
• RAS is now owner or DLUS for the dependent LUs.
• The dependent LU sessions are now same-domain instead of cross-domain.
• RAS has run the configuration services exit with its own copy of the MODELS
major node.
When a VTAM (such as an interchange node or an migration data host) has both
subarea and APPN functions, it has the choice of using either type of network
when called upon to set up a session. The target resource may be accessible
via either subarea or APPN, in which case the installation may wish to influence
the choice. Requirements such as HPR or DLUR will dictate an APPN path,
whereas other factors such as NCP loading, BSC 3270 support or MLTGs may
make a subarea path desirable or even essential for a particular session. The
choice of routing depends on a combination of start options, definitions and
VTAM′s knowledge (acquired dynamically or by definition) of a resource. VTAM
identifies the location of a resource in one of three ways:
• Predefinition
• Incoming search request with the resource as OLU (the resource has
requested a session with an LU on this VTAM)
• Incoming search reply with the resource as DLU (this VTAM has performed a
successful search for the resource)
Thus the installation can influence the choice of subarea or APPN routing either
by predefining resources (which is not normally desirable) or by adjusting
VTAM′s searching algorithms. Since those algorithms are completely different in
subarea and APPN networks, the combination of them in a mixed network is
quite complex. However, an understanding of that combination is essential if
trouble-free migration of a medium or large network is to be achieved.
When a CDRSC is created, it has one or more of the following entries depending
on how it was created:
The complete VTAM subarea search algorithm is shown below. VTAM sends the
CDINIT or DSRLST to each partner VTAM in turn until the resource is found. No
search is issued until a response has been received to the previous one. It is
not permitted in the subarea network to have multiple searches outstanding
concurrently for the same resource. For clarity, we show the
SSCPORD=DEFINED and SSCPORD=PRIORITY cases separately:
Thus a LEN resource can be easily told apart from a subarea CDRSC in the
VTAM display.
When VTAM is capable of both APPN and LEN, it maintains two distinct identities
to cater for the two networks. Its subarea side regards APPN resources exactly
as LEN resources, in other words as independent LUs and therefore CDRSCs. In
APPN, however, resources are not usually predefined and therefore will be
represented as dynamic CDRSCs, which last only for the duration of a session.
An APPN resource is represented in the subarea side of VTAM as a CDRSC with:
• This VTAM shown as the owner (real and/or coded).
• A dummy adjacent link station, ISTAPNPU. This tells VTAM not to try to send
a BIND across a link station, but to pass a search request to its APPN side
so that the APPN network can be searched.
So now we can see that a remote resource can be represented to VTAM as any
or all of subarea, LEN and APPN. It could have (for example) another VTAM as
real owner, this VTAM as coded owner, and a list of link stations among which is
ISTAPNPU. VTAM knows whether a LEN link is active or not because it owns the
link; thus it can ignore the LEN representation of a resource if the LEN
connection is not available. Assuming any LEN link stations are active, this is
how VTAM decides between the three possible options:
• If the choice is between LEN and subarea, LEN takes priority.
• If the choice is between LEN and APPN, APPN takes priority unless
ALSREQ=YES has been coded on the CDRSC definition. This prohibits the
dynamic addition of link stations, including ISTAPNPU, to the CDRSC.
• If the choice is between APPN and subarea, the algorithms described in 9.3,
“How VTAM Combines the Search Algorithms” on page 155 are used.
• If all three representations are known to VTAM, APPN takes priority over LEN
and subarea is suppressed, unless the search has come from the APPN
network. In this case it is too late to reroute it back to APPN (because of the
search algorithms shown in 9.3, “How VTAM Combines the Search
Algorithms” on page 155), so LEN takes priority over subarea.
APPN Search
1. APPN directory. If an entry is found here a directed Locate is sent to verify
that the resource is still in its expected location. If the expected location is
an EN served by this NN, the directed Locate is sent directly to the CP. If the
expected location is elsewhere, it is sent to the NN (owner or server) which
serves the target LU.
2. APPN topology database. If the target is actually a network node CP, it will
be found here and no verification is necessary. VTAM, however, still sends a
directed Locate to verify the target.
3. Domain broadcast. The NN sends a Locate request to all its served end
nodes that have permitted themselves to be searched on a domain
broadcast. Some end nodes, VTAM included, do not permit such a search;
they register all their resources to their NNS thus eliminating unnecessary
searching overhead. There are exceptions, such as the case where a VTAM
is managing an APPN peripheral border. Here VTAM pretends to be an end
node but cannot possibly register all the resources on the far side of the
border; therefore it allows itself to be searched.
4. Full APPN network search. Exactly what this entails depends on the role the
NN is playing in the search and the presence of CDSs in the network. If
there is no CDS then the full search comprises:
a. Full network broadcast. The NN sends a Locate request to all adjacent
NNs with which it has CP-CP sessions. They in turn forward the request
The above algorithm assumes that the NN is performing the search as a result of
being NNS(OLU); in other words it has originated the search. If it is searching as
a result of receiving a broadcast it does not perform the “full APPN search”
actions because some other node does that. If it is searching as a result of
being the entry border node (in other words, it has received the search from a
cross-border partner), then it goes straight into its adjacent cluster tables and
does what they tell it to do. Only if it finds its own name in the tables does it
search its own network, as if it were NNS(OLU).
When VTAM, acting as an ICN or an MDH, is called upon to search for a target
resource it uses all the above definitions, plus its dynamically acquired
knowledge of the resource, to conduct the search. The algorithms it uses are
shown below; for clarity they have been split into several distinct cases because
some options affect it more significantly than others.
Regardless of the setting of all the start options and definitions, if ADJLIST is
coded on a CDRSC definition then that list, and only that list, is used to find the
resource. The start options are ignored. ADJLIST is a good way (indeed, the
only way) of overriding SORDER=APPNFRST if a subarea path is desired to a
given resource. An ADJLIST can include ISTAPNCP if so desired.
You should also be aware that the Directory Services Management Exit can
suppress some of the search steps, but it cannot change the order of the steps.
The best options for your network will vary according to the stage of migration
you have reached, but in general:
• Start with SORDER=ADJSSCP or SUBAREA (for the simpler networks) and
move to APPN or APPNFRST as you migrate. Use ADJSSCP when the
location of resources depends mainly on the NetID, or when the resources
VTAM has a function called CNN routing to optimize the selection of entry and
exit TGs (meaning the PLU side and the SLU side) when an APPN route crosses
a CNN. CNN routing has the following characteristics:
• The APPN rule that forces the lowest-weight route to be chosen is inviolable.
CNN routing only comes into play if the alternative paths across a CNN have
the same weight.
• VTAM nodes propagate subarea numbers in the TG descriptors (CV46) sent
on topology updates and search requests. A CV46 describing a TG that
originates in a subarea node has the subarea number appended in subvector
80. This means that any VTAM in an APPN network can determine if there is
a zero-hop path across any CNN, provided that it has received the right
CV46s with the subarea numbers. VTAM is capable of working out that a TG
from a served end node connects to the same subarea number as the same
TG to the end node, but non-VTAM nodes do not understand such things.
The algorithm that VTAM uses for CNN routing has changed over the years, as
customers have found new configurations that the old rules did not address very
well. At the time of writing (after APAR OW26906 to VTAM V4R4), the algorithm
is as follows:
1. First, VTAM will favor HPR TGs over non-HPR TGs. However, this only
applies if the HPR-capable TG connects an RTP-capable end node to an
ANR-capable node that is not its RTP partner. In other words:
• If the PLU is on an end node and the first two hops of the path are
HPR-capable, then an HPR TG is preferred for the first hop from the PLU.
• If the SLU is on an end node and the last two hops of the path are
HPR-capable, then an HPR TG is preferred for the last hop to the SLU.
2. Next, VTAM will attempt to select entry and exit TGs that are connected to
the same subarea. If the subarea information is available (from the CV46s or
because VTAM owns the partner LU), it will try to match subarea numbers to
find a zero-hop route through the CNN. Note that VR-TGs are assigned a
subarea number of zero, so a VR-TG will never match another TG.
3. If no subarea match is found, and if the HPRNCPBF start option is set to YES,
VTAM will favor ANNC TGs over other TGs. HPRNCPBF allows VTAM to
favor an HPR path for an NCP-attached LU, even if that path would traverse
the same NCP twice (first from NCP to VTAM to reach the RTP endpoint, then
back to the NCP over an HPR connection). Favoring ANNC over NCP TGs
minimizes the risk of such a path.
HPRNCPBF was introduced by APAR OW25950.
4. Next, VTAM will look for a VR-TG if a suitable one is available.
5. If there is no suitable VR-TG, VTAM will pick any other available TG.
A recent refinement has been introduced by APAR OW29969. This extends the
subarea matching concept so that VTAM can determine that two subareas within
a CNN are “closely linked”, rather than the same. The installation must define a
subarea mapping table (VBUILD TYPE=SAMAP) to tell VTAM which subarea
pairs are considered suitable as entry and exit points for sessions. This subarea
mapping method also eliminates the situation where the chosen subarea pair is
so unsuitable that the installation has not even defined a VR between them; in
such a case the session would fail unless VTAM is told to favor a different pair of
SAMAP is used only for route calculation, not for BIND rerouting. Its main
purpose is to allow customers with remote NCPs to optimize routes across what
could be quite a complex CNN. For example, consider the network shown in
Figure 122.
┌────────┐
│ │
┌─┤ EN ├─┐
TG 21 │ │ │ │ TG 22
│ └────────┘ │
│ │
┌────────┐ ┌───┴────┐ ┌──┴─────┐ ┌────────┐
│ │ │ NCP │ │ NCP │ │ NCP │
│ VTAM ├────┤ SA 67 ├────┤ SA 55 ├────┤ SA 8 │
│ │ │ │ │ │ │ │
└────────┘ └───┬────┘ └───┬────┘ └───┬────┘
│ │ │
│ │ │
LUA ┌───┴────┐ LUC
│ NCP │
│ SA 10 │
│ │
└───┬────┘
│
│
LUB
In this diagram, a session from the end node to LUA will always use TG 21,
because VTAM can match LUA′s subarea (67) against the subareas which
connect the CNN to the end node (67 and 55). However, VTAM cannot match the
subareas for LUB and LUC (10 and 8 respectively) against 67 and 55. Therefore,
sessions from the EN to LUB or LUC will use TG 21 and TG22 alternately. TG 21
is clearly less desirable.
The SAMAP table allows you to associate SA55 with SA8 and SA10, so that
VTAM knows that SA55 is a closer match to SA8 and SA10 than SA67 is. Thus
sessions to LUB or LUC which have a choice between TG 21 and TG 22 will be
allocated to TG 22 (SA 55) rather than TG 21 (SA 67). A subarea mapping table
for this purpose might appear as in Figure 123 (note the prefix SA before the
subarea numbers).
VBUILD TYPE=SAMAP
SA10 MAPSTO SUBAREA=SA55
SA8 MAPSTO SUBAREA=SA55
The figure shows two composite network nodes, RAS and RAA, connected
together by both an ANNC TG (21) and a VR-TG. A and B are APPN end nodes
connected to the subarea network by two TGs each. Let us suppose that node B
wishes to establish a session with node A, and that all the TGs have equal
weight for the session COS. RAS, as B′s NN server, must calculate the route.
If every node in the network is RTP-capable, then RAS tries subarea matching as
follows:
1. The first intermediate node on the session path is itself, RAS. The entry
subareas are 28 (TG21) and 8 (TG22). The exit subareas are 28 (TG21) and 0
(VR-TG). Therefore, there is a match between TG21 inbound and TG21
outbound. Both TG21s are chosen.
2. Now RAS must choose the route through RAA. It should have the subarea
numbers available from the topology database and from RAA′s response to
the Locate, so it tries subarea matching again. This time it can see that the
possible exit subareas are 10 (TG21) and 5 (TG22). Only the first matches
the known entry TG (TG21, which goes to subarea 10). Therefore, the route
chosen through RAA is again TG21 - TG21.
This may or may not be the preferred route, but it is probably better than what it
would have been before APAR OW26906. Previously, a VR-TG was associated
with the subarea numbers of its partner VTAMs. Thus subarea matching in this
example would have chosen the route:
Node B - TG21 - RAS - VR-TG - RAA - TG21 - Node A
which passes through all six nodes.
If the OLU is to be the PLU, the OLU is expected to supply the mode. If it does
not do so, a mode name comprising all blanks is sent into the network, and
SSCP(SLU) must then resolve it to a named mode from the SLU definitions. If
the SLU is a dynamic CDRSC with no definitions, VTAM uses the DYNDLGMD
and DYNMODTB start options to determine the correct mode to use.
If the subarea network includes SNI boundaries, some of this work must be done
by proxy since VRs are terminated at SNI boundaries:
• The subarea COS is translated, if necessary, at SNI boundaries as it flows in
the secondary to primary direction.
• The VR list is selected by the gateway SSCP of the PLU at each SNI
boundary.
In the first PLU-initiated example, the PLU as OLU has supplied a mode name.
Since it is the responsibility of SSCP(SLU) to resolve the mode name to a COS,
the mode name is transmitted to VTAM2 on the CDINIT request. VTAM2 duly
resolves the COS, and sends the COS name back to VTAM1 on the CDINIT
response. VTAM 1 needs the COS name to create a VR list for route selection.
In the second example, the OLU has failed to supply a mode name so VTAM1
transmits a blank mode to VTAM2 on the CDINIT request. VTAM2 must now
choose a mode (normally from the SLU′s definitions) and immediately resolves
that mode into a COS. As before, VTAM2 sends the COS name back to VTAM1,
but this time it sends the mode name as well.
The calculation of the route is the responsibility of the network node server of
the PLU. It uses the supplied APPN class of service to assign weights to all the
possible TGs and nodes on the session path, and chooses the path with the
lowest overall weight.
Figure 126 on page 170 shows how this works in the SLU-initiated and
PLU-initiated cases.
The APPN example shown here is the simplest one, that of session initiation
between an end node and its network node server. Any other configuration
would be more complex, but the exchange of mode and COS names between
PLU and SLU is essentially the same.
In the PLU-initiated case, EN1 as CP(OLU) chooses both the mode and the
APPNCOS. Both pieces of information are sent on the Locate request, which in
this case never gets beyond NN2 which is both CP(SLU) and NNS(PLU). As
NNS(PLU), NN2 has all the information it needs to calculate the route, so it
responds with the session RSCV on the Locate reply. EN1 can then start the
session.
NN2 can do all these things without consulting any other node in the network, so
it sends a Locate to EN1 with the mode, the APPNCOS and the session RSCV.
EN1 can start the session at once, and the Locate reply contains a “session
started” indicator to tell NN2 that the BIND has been sent.
Note that in the PLU-initiated case EN1 and NN2 could be any APPN products,
but in the SLU-initiated case both must support session services extensions
(today, that means both must be VTAM). The “session started” flow is part of
the session services extensions APPN option. To support SLU-initiated sessions
The traditional way for VTAM to choose a class of service for a session is by
looking up the mode entry in the mode table associated with the SLU and
reading the COS name from the entry. If the entry is not found in the SLU′ s
mode table, VTAM will look for it in the default table ISTINCLM. Today, both
subarea and APPN COS names can be specified in a mode entry, and indeed the
IBM-supplied mode table ISTINCLM includes the architected APPN COS names
in the appropriate mode entries. In today′s environment, this method is still
used:
• When the second method (COS mapping) will not work because there are no
suitable COS map definitions.
• When this VTAM is the first to see this session request, thus there is no
existing COS to map. The first COS to be resolved on a session is always a
subarea COS, and is always resolved by the mode table lookup method.
If the mode for the session is known then the matter is straightforward. If the
mode name is not recognized then there is a default mode, ISTCOSDF, that
VTAM will use to look up the COS. The use of ISTCOSDF is controlled by the
start option ISTCOSDF, which tells VTAM what kinds of resource may use the
ISTCOSDF mode. As with the “real” mode entry, if VTAM does not find
ISTCOSDF in the table associated with the SLU it will look for it in ISTINCLM.
Figure 127 on page 172 shows part of a mode table with the appropriate COS
entries for subarea (COS=) and APPN (APPNCOS=) coded. The IBM-supplied
table ISTINCLM now includes APPNCOS operands; it has never had subarea
COS names defined.
The second method of COS resolution is for VTAM to translate a COS supplied
on one type of network to a different COS on the other type of network. This
involves the coding of COS mapping tables in VTAMLST and their activation.
The COS mapping method is preferred by VTAM, even if the appropriate COS
entries are in all the mode tables.
Figure 128 illustrates sample COS mapping tables from subarea to APPN
(SATOAPPN) and APPN to subarea (APPNTOSA).
If you wish to specify the default subarea COS (blank) as either the origin or the
target of a COS mapping table entry, you need to install APAR OW28568. This
function was introduced after the availability of VTAM V4R4. If you do not have
OW28568 applied, you can still map the blank subarea COS to an APPN COS by
means of the DEFAULT entry, which VTAM uses if it cannot find the incoming
subarea COS in the table. Without OW28568 it is not possible to map an APPN
COS to the blank subarea COS.
However, if the route being established is for an RTP connection crossing from
APPN to subarea, VTAM will never use the mode table lookup method to
determine the subarea COS. This is because a route setup (as opposed to a
BIND) may not contain a mode name. In this case VTAM will first try the
APPNTOSA mapping method, and if that fails it will choose the default ISTVTCOS
class of service in the subarea portion of the network.
In the diagram, the four nodes are all VTAM nodes. If EN or NN were not VTAM,
then an SLU-initiated session would not be possible because only VTAM today
supports session services extensions.
The first thing to realize here is that there are three subarea networks in the
diagram. Each of the stand-alone VTAMs (EN and NN) uses subarea protocols
between the session partner LU and the boundary function to the adjacent APPN
network. Because OLU=SLU, the subarea and APPN protocols fit together quite
well, as follows:
1. The SLU issues a logon (or INIT_SELF) request to its owning VTAM. As
SSCP(OLU), the VTAM network node determines the session mode to be
used. Since it is also SSCP(SLU), it immediately resolves the mode into a
subarea class of service (COS) and sends the session request on its way.
From the above discussion, you can see that SLU-initiated sessions in a mixed
network work quite well. The mode for the whole network is chosen by the same
node (OLU=SLU), the subarea and APPN classes of service are chosen on the
same flows, and the COS mapping works as you would expect it to work.
Note also that if a VR-TG is defined across the central subarea network, the
APPN Locates flow across the VR-TG and no subarea-to-APPN protocol
conversions are needed. Such a route would be calculated by NNS(PLU).
Figure 130 shows how the same network deals with a PLU-initiated session.
Once again subarea protocols are used in three networks and APPN protocols in
two networks. This time, however, they do not fit together as smoothly as in the
SLU-initiated example:
1. The PLU issues a session request (OPNDST, probably) to its owning VTAM,
with a mode specified. This VTAM EN is CP(OLU) for the session, so it
selects an APPN COS. The architecture allows it to leave COS selection to
its NNS, but VTAM always chooses the COS itself. It tries to do this from the
SLU definitions, but these probably do not exist since the SLU is out in the
APPN network and is (or will be) represented by a dynamic CDRSC. If this is
so, VTAM picks the mode table specified in the DYNMODTB start option if it
exists; otherwise, it picks ISTINCLM.
VTAM chooses an APPN COS as follows:
a. It takes the chosen mode from the chosen mode table and picks out the
subarea and APPN COS values.
b. It tries to use the SATOAPPN COS mapping table to convert the subarea
COS into an APPN COS.
In the PLU-initiated case the APPN and subarea classes of service are chosen
on different flows: the APPN in the PLU-to-SLU direction and the subarea in the
SLU-to-PLU direction. The COS mapping still works, but it does not work in the
obvious way. Moreover, if the PLU exercises its right not to provide a mode
name, the flows are complicated still further:
1. The VTAM EN, when choosing an APPN COS for the session, will try to
resolve it from the DLOGMOD in the SLU′s definitions, or the first entry in
the mode table if there is no DLOGMOD. However, it is likely that the SLU is
a dynamic CDRSC and therefore it will try to use:
a. DYNMODTB, then ISTINCLM, for the mode table
b. DYNDLGMD, then the first entry in the table, for the mode
Once the correct mode entry has been determined, the method of choosing
the APPN COS is the same as in the case where the PLU supplied a mode.
However, the mode name forwarded into the APPN network is still blank.
2. The ICN at the PLU side of the subarea network forwards the blank mode in
the CDINIT to the subarea network.
3. The ICN at the SLU side of the subarea network must again resolve a blank
mode to an APPN COS. As with CP(OLU), it must contend with both a blank
mode and a dynamic CDRSC for the SLU.
4. CP(SLU) receives the blank mode, and its subarea side resolves it to a real
mode. At last the real SLU is available, probably with suitable definitions,
and this VTAM finally selects the real mode required for the session. It is
this real mode name that is now sent back into the network.
5. The ICN at the SLU side now selects a subarea COS as before, but this time
there is a possible complication. It first tries to use the COS mapping
method, which is based on the APPN COS chosen from the blank mode. If
this fails, however, it will attempt the table lookup method for which it has
the real mode. Thus even if the real mode is available, the subarea COS
chosen could be the result of the blank mode.
6. The ICN at the PLU side forwards the new mode to the EN, as before.
7. CP(PLU) receives a new mode name (which of course differs from the one it
chose on the outbound flow to resolve the APPN COS). The new mode is
given to the PLU to start the session. As before, CP(PLU) performs an
APPN-to-subarea COS translation which may originate (as with ICN(SLU))
from the blank mode or the real mode depending on circumstances.
Thus it is perfectly possible to end up with a different subarea COS to the one
you started with. The translation from subarea to APPN to subarea occurs at the
same node in different directions, and may well not yield the answer you want.
Note:
The APPN routes in this example are calculated by NNS(PLU) and ICN(PLU), in
the left- and right-hand networks respectively. Again there are IC-TGs in both
One of the main reasons for customers migrating to APPN is the implementation
of a sysplex. The Parallel Sysplex provides some unique advantages to the
VTAM installation, namely generic resources and multinode persistent sessions.
Generic resources allows multiple application copies on a sysplex to present a
single image to the user, resulting in better performance through load balancing
and improved availability. The multinode persistent sessions function enables
sessions to survive, without disruption, the failure of a VTAM, MVS or even a
processor in a sysplex. The redbook VTAM in a Parallel Sysplex Environment ,
SG24-2113, describes the workings of SNA in a sysplex.
The generic resources and multinode persistent sessions (MNPS) functions both
require APPN to be implemented within the sysplex; indeed, MNPS depends on
HPR for its operation. This means that all the VTAMs within the sysplex must be
able to establish CP-CP sessions with each other. For the purpose of generic
resources VR-TG connections are sufficient, but if you want MNPS to work
reliably you need FID2 connections instead (or as well).
VTAM systems within a sysplex have two primary means of communicating with
each other. They may use dedicated ESCON channel-to-channel connections, or
they may use the MVS Cross-System Coupling Facility (XCF). Both offer
high-performance APPN connectivity, but whereas a dedicated
channel-to-channel link can use base APPN or HPR, XCF requires HPR.
In the earlier sections of this redbook, we have described the basic migration
scenario as being:
1. Convert CMC(s) to APPN
2. Convert boundary links (former LEN connections) to APPN
3. Convert subarea links (between VTAMs) to APPN, node by node
This plan holds good for networks both small and large, but the installation of a
sysplex may require a somewhat different approach. Here we are adding a
whole new APPN network to what might be a pure subarea network, and we
have to integrate them.
The CMC and the backup CMC (BCMC) are outside the sysplex, which contains
NNs and ENs. The NNs must be ICNs because there is no APPN capability
outside the sysplex; the ENs are MDHs to retain subarea capability. All VTAMs
and NCPs are connected to each other using ESCON channels as shown; the
VTAMs in the sysplex may or may not also have XCF connections. Inside the
sysplex the CP-CP sessions are on FID2 links. All VTAMs have SSCP sessions
with each other (as they must), over subarea MPC connections. All VTAMs
inside the sysplex are connected by VR-TGs (not necessarily fully meshed), to
avoid parallel APPN and subarea paths. The CMC owns all the NCPs as its
name implies.
The difference here is that there are no subarea connections, or SSCP sessions,
to the end nodes within the sysplex. The physical connectivity is the same, but
Here there are CP-CP sessions over FID2 links within and outside the sysplex.
To complete the picture (and to ensure that independent parallel subarea and
APPN paths do not exist) a VR-TG is defined between the two CMCs. The whole
of the network is now APPN, with the subarea portion confined to the two CMCs
and their NCP backbone. The NNs within the sysplex are now solely NNs, not
ICNs.
Now every node has a subarea path to every other node, and all the VTAMs
have both SSCP sessions with each other, and CP-CP sessions where required.
The only FID2 connections are within the sysplex (ANNC or XCF). All NNs are
The disadvantage is simply that, because the new portion of the network is
subarea capable, a full complement of CDRM, PATH, adjacent SSCP and CDRSC
statements will be required.
This is because when you convert a VTAM-based network to APPN, the same
management information flows between the same nodes to a very large extent.
The connections may be APPN and the sessions may be LU 6.2, rather than
subarea and SSCP-PU/LU, but the flow of data is the same. Whether you use
DLUR/DLUS for your dependent LU sessions or not, you still have SSCP-PU and
SSCP-LU sessions to carry the same management data on their behalf. The
biggest change is the ability to “see” topology and session information beyond
what were once LEN connections, and start managing a wider network.
The format and contents of the data flowing on SSCP-PU sessions depends on
the level of currency of the node containing the PU. The more up-to-date nodes
use network management vector transport (NMVT) request units containing
management vectors. Older nodes send RECMS and RECFMS request units.
The issue with APPN, of course, is that there are no PUs and therefore no
SSCP-PU sessions. Ownership is distributed or nonexistent, so a new structure
is needed which defines a relationship between a managing node and a
MDS support has been available in NetView since Version 2 Release 2 (and is
therefore in the TME10 NetView product). Since Version 4 Release 1 VTAM has
also provided MDS support; NetView from V2R4 onwards can be configured to
use either VTAM′s or its own support. MDS is also supported by current levels
of the Communications Server suite, 3746, 2216, 2210, AS/400 and 3174.
In this section we describe how these functions are performed, and how they
appear, in an APPN network.
The DISPLAY command gives a view of connectivity and status in both subarea
and APPN networks. Because the APPN support in VTAM is built on the
independent LU support introduced in V3R2, if you understand the LEN/ILU
displays you will almost certainly have no difficulty with the APPN displays.
Figure 33 on page 72 and Figure 54 on page 85 are annotated illustrations of
typical resource displays in an APPN environment.
The VARY command is used in exactly the same way in APPN, except that you
have to remember that VTAM may no longer own the same resources as it did in
the subarea case. Activation and inactivation commands control only direct
connections (adjacent link stations) and DLUR resources.
Hardware monitor alerts, events and statistics which originate within the
(subarea) domain of this VTAM will be gathered the same way (SSCP-PU
sessions), and displayed in the same way, as before. Hardware monitor data
which originates elsewhere will be displayed the same way, but transported:
• By MDS, if from an APPN-connected node
• By SSCP-PU sessions, if from a DLUR resource
• By SSCP-PU sessions, if from an adjacent type 2 / type 2.1 mixed node
• By MDS, as previously in the subarea network, if from another NetView
which gathered the data in the first place.
The NetView Session Monitor will be aware (as before) of sessions passing
through this VTAM or its owned NCPs (its subarea domain), since VTAM will still
have been involved in the session setup. If the session is wholly or partly APPN
then the Session Monitor will have additional APPN route information at its
disposal:
• The Session Configuration panel shows APPN COS and transmission priority.
• A subsequent new panel shows the APPN route in terms of CP names and
TG numbers.
Note, however, that this does not apply to HPR sessions using VTAM as an ANR
node. The HPR design ensures that ANR nodes are not aware of sessions or
RTP pipes passing through them.
Response Time Monitor (RTM) data is still available in an APPN network, since
the SSCP-PU sessions required for it are still available, albeit perhaps
encapsulated in a DLUR/S connection.
The 3174 Central Site Control Facility (CSCF) depends on SSCP-PU sessions;
therefore if a CSCF-controlled 3174 is separated from its owning subarea VTAM
it must utilize DLUR/S for CSCF operation. CSCF is the function that allows a
NetView operator to manage a 3174 as if from the operator panel on a directly
attached terminal.
None of these requires SSCP-PU sessions; therefore they all work the same way
in APPN as in subarea networking. In addition, NPM can collect performance
data from an LU within a 3746-9X0 NNP. This LU is dependent, so it utilizes the
DLUR support in the 3746 to establish the required session.
One other product that provides new function in APPN is the NetView Remote
Operations Manager. This enables the NetView operator to issue commands to
agent programs running on remote AS/400 machines. Remote Operations
Manager uses MDS Operations Management protocols over LU 6.2 sessions to
transport the data.
Interchange Node HOSTSA= network yes yes, NCP/VTAM yes, for subarea
(ICN) NODETYPE=NN node become a CNN connections
Migration Data HOSTSA= end node data host no (see note) yes, for subarea
Host (MDH) NODETYPE=EN only connections
Note: A VTAM APPN-only node cannot activate or own resources in an NCP. It can establish a type 2.1 connection with the
NCP, allowing it to connect to APPN nodes via the NCP. NCP must be at Version 6 Release 2 or later to support APPN links.
SSCPORD The old subarea option, but it affects the way the N PRIORITY,
APPN/subarea search patterns work. See DEFINED
Chapter 9, “Searching and Routing in a Mixed
Subarea/APPN Network” on page 151.
XNETALS Used to control the attachment of adjacent nodes N YES (APPN), NO (others),
with different NetIDs to VTAM. The default for YES/NO
APPN connections is YES, which is required for
an APPN border. Can be overridden on the PU
definition.
Note: All defaults are highlighted. Dynamic means modifiable by F proc,VTAMOPTS except in the case of the TRACE
option for which F proc,TRACE is used.
Please see VTAM Operation , SC31-8567, for a complete list of options and
syntax.
Notes:
1. NSEARCH will cause a network search to be performed for the resource.
2. ″E″ is the abbreviation for SCOPE=ALL, N for SCOPE=ONLY. Many
commands default to SCOPE=ONLY. Specifying ″E″ will give much more
information.
3. IDTYPE covers many more areas, APPN types are the only ones listed here.
Depending on the VTAM DSPLYWLD start option, wildcard values can be
used for the resource name. An asterisk can be specified for the NETID
portion of the ID.
4. SCOPE can be used to limit the display. The options are as follows:
SCOPE=ALL|ACT|ACTONLY|CONCT|INACT|INACTONLY|ONLY|PENDING|RESET
We have two VTAM hosts in this initial configuration: RAA (subarea 10) and RAS
(subarea 28). Both are connected using a CTC and later an MPC in preparation
for the APPN migration.
There are two NCPs, RA5NCPX (subarea 5) and RA8NCPX (subarea 8).
RA5NCPX is channel-attached to RAA and link-attached to RA8NCPX. RA8NCPX
is channel-attached to RAS and link-attached to RA5NCPX. Both NCPs are
loaded and owned from RAA. This makes up the CMC.
Both NCPs are attached to a token-ring and have connectivity to each other
through the ring.
ATCSTR10
*******************************************************************
* *
* ATCSTR10 FOR RAA *
* THIS IS SUBAREA 10 STAGE 1 ; SUBAREA *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
CONFIG=10,
CSALIMIT=0,
DYNLU=NO,
HOSTPU=ISTPUSA0,
HOSTSA=10,
IOBUF=(100,384,5,,1,30),
NETID=USIBMRA,
NOTRACE,TYPE=VTAM,IOINT=0,
SSCPID=99,
SSCPNAME=RAA,
SUPP=NOSUP,
TNSTAT,CNSL,
XNETALS=YES
ATCCON10
RAAATSO, TSO APPL
RAAANETV, NETVIEW V2.4 APPL
RAAPATH, SA10 PATH TABLE
RAACDRM, CDRM
RAACDRSC, CDRSCS
RAAADJ, ADJSSCP TABLE
RAAMPC1, CTC TO SA 28 (FID4 MPC)
RAALCL1, LOCAL SNA 3174-11L
RAASWMN1, SW-TR MN FOR 3174-63R
RAASWMN2, SW-TR MN FOR CM/2 (KEN)
RAASWMN3, SW-TR MN FOR CM/2 (ANAR)
RAASWMN4, SW-TR MN FOR CM/2 (RABINDER)
RAASWMN5, SW-TR MN FOR AS/400
RA5NCPX, NCP SA 5
RA8NCPX REMOTE NCP SA8
RAAPATH
PATH DESTSA=5,
ER0=(5,1),ER3=(5,1),
VR0=0,
VRPWS00=(10,30),VRPWS01=(10,30),VRPWS02=(10,30)
PATH DESTSA=8,
ER0=(8,1),ER1=(28,1),ER2=(5,1),
ER3=(5,1),
VR0=0,
VRPWS00=(10,30),VRPWS01=(10,30),VRPWS02=(10,30),
VR1=3,
VRPWS10=(20,60),VRPWS11=(20,60),VRPWS12=(20,60),
VR2=2,
VRPWS20=(20,60),VRPWS21=(20,60),VRPWS22=(20,60),
VR3=1,
VRPWS30=(20,60),VRPWS31=(20,60),VRPWS32=(20,60)
PATH DESTSA=20,
ER0=(20,1),ER2=(20,1),
VR0=0,
VRPWS00=(10,30),VRPWS01=(10,30),VRPWS02=(10,30)
PATH DESTSA=28,
ER0=(20,1),ER1=(5,1),ER2=(5,1),
ER3=(28,1),ER4=(28,1),
VR0=0,
VRPWS00=(20,60),VRPWS01=(20,60),VRPWS02=(20,60),
VR1=1,
VRPWS10=(30,90),VRPWS11=(30,90),VRPWS12=(30,90),
VR2=2,
VRPWS20=(30,90),VRPWS21=(30,90),VRPWS22=(30,90),
VR3=4,
VRPWS30=(10,30),VRPWS31=(10,30),VRPWS32=(10,30)
RAAADJ
*******************************************************************
* *
* ADJSSCP TABLE FOR RAA *
* THIS IS SUBAREA 10 PRIOR TO MIGRATION FOR RED BOOK *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
* RAS SUBAREA 28 IS THE ONLY OTHER VTAM IN THE NETWORK *
* *
*******************************************************************
VBUILD TYPE=ADJSSCP
RAS ADJCDRM
RAACDRM
*******************************************************************
* *
* CDRM MAJORNODE FOR RAA *
* THIS IS SUBAREA 10 PRIOR TO MIGRATION FOR RED BOOK *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
VBUILD TYPE=CDRM
NETWORK NETID=USIBMRA
RAA CDRM SUBAREA=10,CDRDYN=YES,CDRSC=OPT
RAK CDRM SUBAREA=20,CDRDYN=YES,CDRSC=OPT
RAS CDRM SUBAREA=28,CDRDYN=YES,CDRSC=OPT
RAACDRSC
VBUILD TYPE=CDRSC
NETWORK NETID=USIBMRA
RASAN CDRSC CDRM=RAS
RASANLUC CDRSC CDRM=RAS
RASAT CDRSC CDRM=RAS
RAALCL1
*******************************************************************
* *
* LCL MAJORNODE FOR RAA FOR SNA LOCAL ATTACHED 3174-11L. *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
VBUILD TYPE=LOCAL
RAALCLP PU CUADDR=E40,PUTYPE=2,MAXBFRU=10, *
MODETAB=AMODETAB,DLOGMOD=M2SDLCQ,USSTAB=US327X, *
CONNTYPE=LEN,XID=YES, *
VPACING=0
RAALCL02 LU LOCADDR=2
RAALCL03 LU LOCADDR=3
RAALCL04 LU LOCADDR=4
RAALCL05 LU LOCADDR=5
RAALCL06 LU LOCADDR=6
* *
*******************************************************************
RAAMPC1
*******************************************************************
* *
* MPC MAJORNODE FOR RAA(10) TO RAS(28) - FOR FID4 MPC CONNECTION *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
RAACRAS VBUILD TYPE=CA
RAAGMP1 GROUP LNCTL=MPC,ISTATUS=ACTIVE,MAXBFRU=3
RAALMP1 LINE READ=(C12),WRITE=(C14)
RAAPMP1 PU PUTYPE=4,TGN=1
The following definitions are for the token-ring-attached 3174-63R with MAC
address 400031740004 and IDBLK=017 and IDNUM=31744.
RAASWMN1
*******************************************************************
* *
* SWMN MAJORNODE FOR RAA FOR TOKEN-RING ATTACHED 3174-63R. *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
VBUILD TYPE=SWNET,
MAXNO=1,
MAXGRP=1
RAATRPU1 PU ADDR=C1,
IDBLK=017,
IDNUM=31744,
MAXOUT=7,
MAXPATH=1,
PASSLIM=7,
SSCPFM=USSSCS, (V) VTAM
USSTAB=US327X, (V) VTAM
MODETAB=ISTINCLM,
DLOGMOD=D4C32XX3, (V) VTAM
PUTYPE=2
RAATRPH1 PATH DIALNO=0004400031740004,
GRPNM=EG05L01
RAATLU12 LU LOCADDR=2
RAATLU13 LU LOCADDR=3
RAATLU14 LU LOCADDR=4
RAATLU15 LU LOCADDR=5
RAASWMN2
*******************************************************************
* *
* SWMN MAJORNODE FOR RAA FOR T-RING ATTACHED PS/2 77 CM/2 (KEN) *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
VBUILD TYPE=SWNET,
MAXNO=1,
MAXGRP=1
RAATRPU2 PU ADDR=01,
IDBLK=05D,
IDNUM=05150,
MAXOUT=7,
MAXPATH=1,
SSCPFM=USSSCS, (V) VTAM
USSTAB=US327X, (V) VTAM
MODETAB=ISTINCLM,
DLOGMOD=D4C32XX3, (V) VTAM
PASSLIM=7,
PUTYPE=2
RAATRPH2 PATH DIALNO=0004400001050001,
GRPNM=EG05L01
RAATLU22 LU LOCADDR=2
RAATLU23 LU LOCADDR=3
RAATLU24 LU LOCADDR=4
RAATLU25 LU LOCADDR=5
The following definition is for an AS/400. Note that the CPNAME parameter is
used instead of IDBLK and IDNUM. This is a better way of defining type 2.1
nodes to VTAM. IDBLK/IDNUM, however, should be used for DLUR resources
since they should not be confused with their DLUR′s CP. Note also the old way
of coding an independent LU (LOCADDR=0). This will be converted to a CDRSC
by VTAM so you might as well have the extra flexibility and code it as a CDRSC
yourself.
OPTIONS NEWDEFN=(YES,ECHO,NOSUPP),USERGEN=(FNMNDFGN)
*******************************************************************
* RA5NCP 3745-41A CCU A CA1=E28 CA2=E1F CA3=E20 *
* *
* NCP MAJORNODE FOR SA 5 *
* *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
* LICTYPE LINES: LICTYPE LINES: *
* 1 00 3174-63R 1 04 *
* 01 INN(19.2) 05 *
* 02 06 *
* 03 07 *
* *
* NTRI INTERFACES: *
* ================ *
* TICTYPE TR *
* 2 1088 MAC=4000 0105 0000 4MBPS *
* 2 1089 INN (TO SA 8) MAC=4000 0105 0001 4MBPS *
* *
*******************************************************************
*
PRINT NOGEN
* STATOPT=′ SA05 3745′
PCCUPE1F PCCU CUADDR=E28, SA11 WTCOS MVS/SP. VM ADDR E1F *
AUTODMP=YES, ONLY ONE AUTODMP-HOST IF TWINTAIL *
AUTOIPL=NO, ONLY ONE AUTOIPL-HOST IF TWINTAIL *
AUTOSYN=YES, USE THE ALREADY LOADED NCP IF OK *
BACKUP=YES, RESOURCE TAKEOVER PERMITTED *
CHANCON=COND, CONDITIONAL CONTACT REQ. TO NCP SENT*
DUMPDS=NCPDUMP, DUMP DATASET *
MDUMPDS=NCPDMOSS, MOSS DUMP DATASET *
CDUMPDS=NCPDCSP, SCANNER DUMP DATASET *
MAXDATA=5000, *
OWNER=RAA, *
SAVEMOD=YES, *
VFYLM=YES, VERIFY LMOD WHEN LOADING *
SUBAREA=10 HOSTSA VTAM VER 3 MVS/SP
* STATOPT=′ SA05 3745′
PCCUPE20 PCCU CUADDR=E20, SA11 WTCOS MVS/SP. VM ADDR E1F *
AUTODMP=YES, ONLY ONE AUTODMP-HOST IF TWINTAIL *
AUTOIPL=NO, ONLY ONE AUTOIPL-HOST IF TWINTAIL *
AUTOSYN=YES, USE THE ALREADY LOADED NCP IF OK *
BACKUP=YES, RESOURCE TAKEOVER PERMITTED *
CHANCON=COND, CONDITIONAL CONTACT REQ. TO NCP SENT*
DUMPDS=NCPDUMP, DUMP DATASET *
MDUMPDS=NCPDMOSS, MOSS DUMP DATASET *
CDUMPDS=NCPDCSP, SCANNER DUMP DATASET *
MAXDATA=5000, *
OWNER=RAS, *
SAVEMOD=YES, *
VFYLM=YES, VERIFY LMOD WHEN LOADING *
SUBAREA=28 HOSTSA VTAM VER 3 MVS/SP
OPTIONS NEWDEFN=(YES,ECHO,NOSUPP),USERGEN=(FNMNDFGN)
*******************************************************************
* RA8NCP 3745-41A CCU B CA1=E20 *
* *
* NCP MAJORNODE FOR SA 8 (LINK-ATTACHED THROUGH SA 5) *
* *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
* LICTYPE LINES: LICTYPE LINES: *
* 1 32 1 36 *
* 33 CM/2 GW 37 *
* 34 INN (19.2) 38 *
* 35 39 *
* *
* NTRI INTERFACES: *
* ================ *
* TICTYPE TR *
* 2 1092 MAC=4000 0108 0000 4MBPS *
* 2 1093 INN (TO SA 5) MAC=4000 0108 0001 4MBPS *
* *
*******************************************************************
*
PRINT NOGEN
PCCUA010 PCCU SUBAREA=10, *
AUTODMP=YES, ONLY ONE AUTODMP-HOST IF TWINTAIL *
AUTOIPL=NO, ONLY ONE AUTOIPL-HOST IF TWINTAIL *
AUTOSYN=YES, USE THE ALREADY LOADED NCP IF OK *
BACKUP=YES, RESOURCE TAKEOVER PERMITTED *
CHANCON=COND, CONDITIONAL CONTACT REQ. TO NCP SENT*
DUMPDS=NCPDUMP, DUMP DATASET *
MDUMPDS=NCPDMOSS, MOSS DUMP DATASET *
CDUMPDS=NCPDCSP, SCANNER DUMP DATASET *
MAXDATA=5000, *
SAVEMOD=YES, *
RNAME=RA5PRA8, *
LOADSTA=RA5PRA8, *
DUMPSTA=RRA5PRA8, *
OWNER=RAA, *
VFYLM=YES
* STATOPT=′ SA08 3745′
PCCUBE20 PCCU CUADDR=E20, SA11 WTCOS MVS/SP. VM ADDR E1F X
AUTODMP=YES, ONLY ONE AUTODMP-HOST IF TWINTAIL X
AUTOIPL=NO, ONLY ONE AUTOIPL-HOST IF TWINTAIL X
AUTOSYN=YES, USE THE ALREADY LOADED NCP IF OK X
BACKUP=YES, RESOURCE TAKEOVER PERMITTED X
CHANCON=COND, CONDITIONAL CONTACT REQ. TO NCP SENTX
DUMPDS=NCPDUMP, DUMP DATASET X
MDUMPDS=NCPDMOSS, MOSS DUMP DATASET X
CDUMPDS=NCPDCSP, SCANNER DUMP DATASET X
MAXDATA=5000, X
OWNER=RAS, X
VFYLM=YES, VERIFY LMOD WHEN LOADING X
SUBAREA=28 HOSTSA VTAM VER 3 MVS/SP
ATCSTR28
********************************************************
* VTAM START OPTIONS FOR RAS BEFORE MIGRATION TO APPN *
********************************************************
CONFIG=28,
MAXSUBA=255,
CSALIMIT=0,
DYNLU=YES,
DYNASSCP=YES,
GWSSCP=YES,
HOSTPU=ISTPUS28,
HOSTSA=28,
IOBUF=(100,500,5,,1,30),
NCPBUFSZ=2048,
NETID=USIBMRA,
NOTRACE,
TYPE=VTAM,
IOINT=0,
PPOLOG=YES,
SSCPID=28,
SSCPNAME=RAS,
SUPP=NOSUP,
XNETALS=YES
ATCCON28
***************************************************
* CONFIGURATION LIST FOR SA 28 FOR SA NETWORK *
* BEFORE MIGRATION TO APPN *
***************************************************
RASTSO, TSO APPL
RASNETV, NETVIEW APPL
RASLCL1, LOCAL 3270 DEFN
RASPATH, SA28 PATH TABLE
RASCDRM, CDRM TABLE
RASMPC, MPC TO SA10
RASADJ, ADJSSCP TABLE
RASCA8 CA FOR NCP8
RASMPC
*************************************************
* DEFINITION FOR MPC LINK TO RAA (SA10) BEFORE *
* MIGRATION TO APPN NETWORK *
*************************************************
RASMPC VBUILD TYPE=CA
* MPC LINK FROM RAS TO RAA
RASMPCG GROUP LNCTL=MPC
RASMPCL LINE READ=(C14),WRITE=(C12)
RASMPCP PU PUTYPE=4,ISTATUS=ACTIVE
RASLCL1
LBUILD
RAST420 LOCAL CUADDR=420,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST421 LOCAL CUADDR=421,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST422 LOCAL CUADDR=422,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST423 LOCAL CUADDR=423,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST424 LOCAL CUADDR=424,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST425 LOCAL CUADDR=425,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST426 LOCAL CUADDR=426,TERM=3277, X
MODETAB=AMODETAB,USSTAB=US3270, X
ISTATUS=ACTIVE,SPAN=(SPH28L),DLOGMOD=M2BSCNQ
RAST427 LOCAL CUADDR=427,TERM=3286,FEATUR2=(MODEL2), X
MODETAB=AMODETAB,ISTATUS=INACTIVE, X
SPAN=(SPH28L)
RASCA8
*************************************************
* CA MAJNODE FROM RAS TO NCP8 USING ADDRESS E20*
*************************************************
RASCA8 VBUILD TYPE=CA
* CHANNEL LINK FROM 28 TO 8
RASE20G GROUP LNCTL=NCP
RASE20L LINE ADDRESS=E20
RASE20P PU ISTATUS=ACTIVE,PUTYPE=4,MAXDATA=5000
RASADJ
********************************************
* ADJSSCP TABLE FOR RAS TO USIBMSC NETWORK *
* VIA SUBAREA 20 (RAK). RAA IS ANOTHER SSCP*
* IN THIS NETWORK (USIBMRA) *
********************************************
USIBMRA VBUILD TYPE=ADJSSCP
RAA ADJCDRM
USIBMSC NETWORK NETID=USIBMSC
RAA ADJCDRM
RASCDRM
**************************************************
* CDRM MAJNODE FOR RAS BEFORE MIGRATION TO APPN *
**************************************************
VBUILD TYPE=CDRM
NETWORK NETID=USIBMRA
RAS CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=28
RAA CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=10
RASPATH
PATH DESTSA=5,
ER0=(5,1),ER1=(8,1),ER2=(8,1),
ER3=(10,1),
VR0=0,
VRPWS00=(10,30),VRPWS01=(10,30),VRPWS02=(10,30),
VR1=2,
VRPWS10=(20,60),VRPWS11=(20,60),VRPWS12=(20,60),
VR2=1,
VRPWS20=(20,60),VRPWS21=(20,60),VRPWS22=(20,60),
VR3=3,
VRPWS30=(20,60),VRPWS31=(20,60),VRPWS32=(20,60)
PATH DESTSA=8,
ER0=(8,1),ER1=(8,1),ER2=(10,1),
ER3=(10,1),
VR0=0,
VRPWS00=(10,30),VRPWS01=(10,30),VRPWS02=(10,30),,
VR1=3,
VRPWS10=(30,90),VRPWS11=(30,90),VRPWS12=(30,90),
VR2=2,
VRPWS20=(30,90),VRPWS21=(30,90),VRPWS22=(30,90)
PATH DESTSA=10,
ER0=(20,1),ER1=(8,1),ER2=(8,1),
ER3=(10,1),
VR0=0,
VRPWS00=(20,60),VRPWS01=(20,60),VRPWS02=(20,60),
VR1=1,
VRPWS10=(30,90),VRPWS11=(30,90),VRPWS12=(30,90),
VR2=2,
VRPWS20=(30,90),VRPWS21=(30,90),VRPWS22=(30,90),
VR3=3,
VRPWS30=(10,30),VRPWS31=(10,30),VRPWS32=(10,30)
PATH DESTSA=20,
ER0=(20,1),ER1=(20,1),ER2=(10,1),
VR0=0,
VRPWS00=(10,30),VRPWS01=(10,30),VRPWS02=(10,30),
VR1=2,
VRPWS10=(20,60),VRPWS11=(20,60),VRPWS12=(20,60)
***********************************************************************
* *
* MEMBER NAME: IBMTGPS *
* *
* Descriptive name: IBM-Supplied TG Profiles *
* *
* STATUS: ACF/VTAM VERSION 4 RELEASE 3 *
* *
* COPYRIGHT: LICENSED MATERIALS - PROPERTY OF IBM *
* *
* 5695-117 (C) COPYRIGHT IBM CORP. 1992. *
* ALL RIGHTS RESERVED. *
* *
* U.S. GOVERNMENT USERS RESTRICTED RIGHTS - *
* USE, DUPLICATION OR DISCLOSURE RESTRICTED BY *
* GSA ADP SCHEDULE CONTRACT WITH IBM CORP. *
* *
* SEE COPYRIGHT INSTRUCTIONS. *
* $MAC(IBMTGPS),COMP(TRS),PROD(VTAM): IBM Supplied TG Profiles *
***********************************************************************
*
**********************************************************************
* Ethernet *
**********************************************************************
ETHERNET TGP COSTTIME=0,COSTBYTE=0,SECURITY=UNSECURE, *
PDELAY=NEGLIGIB,CAPACITY=10M
*
**********************************************************************
* Token Ring *
**********************************************************************
TOKNRING TGP COSTTIME=0,COSTBYTE=0,SECURITY=UNSECURE, *
PDELAY=NEGLIGIB,CAPACITY=4M
*
**********************************************************************
* ISDN Non-Switched *
**********************************************************************
ISDNNSWT TGP COSTTIME=0,COSTBYTE=0,SECURITY=UNSECURE, *
PDELAY=TERRESTR,CAPACITY=64K
******************************************************************
* *
* VTAM SWITCHED MAJOR NODE FOR ITSC OFFICES *
* *
* CONNTYPE=LEN, WILL ALLOW ILUS COMING IN FROM THIS X
* PU TO HAVE THE PU IN THE ALSLIST FOR THE ILU. X
* *
* CONNTYPE=APPN WILL CAUSE ISTAPNPU TO BE THE ONLY X
* ENTRY IN THE ALSLIST FOR ILUS COMING IN FROM THIS X
* PU EVEN THOUGH THE PU WILL SHOW UP AS THE X
* ADJACENT LINK STATION (BUT NOT IN THE ALSLIST) X
* *
******************************************************************
SWMODEL VBUILD TYPE=MODEL
PUMOD05D PU ADDR=13,
ANS=CONT,
DISCNT=NO,
CONNTYPE=APPN,
MAXDATA=1033,
PUTYPE=2
*
LU6205D LU LOCADDR=0,DLOGMOD=DSIL6MOD
LUMOD05D LU LOCADDR=2,
PACING=0,
DLOGMOD=D4C32XX3,
MODETAB=ISTINCLM,
USSTAB=US327X,
VPACING=0
******************************************************************
* *
* MODEL FOR IDBLK X′017′ - PC 3270 EMULATION *
* *
******************************************************************
PUMOD017 PU ADDR=C1,
ANS=CONT,
PUTYPE=2
LUMOD017 LU LOCADDR=2,
PACING=0,
DLOGMOD=D4C32XX3,
MODETAB=ISTINCLM,
USSTAB=US327X,
VPACING=0
When we migrate VTAM host RAA to an interchange node, with RAS still a
subarea host, the subarea VTAM definitions on RAA are modified. To
summarize:
• On VTAM RAA:
− The start options in ATCSTR10 are changed.
− All other major nodes on RAA remain the same as in the subarea
environment. Refer to D.1, “VTAM RAA Definitions” on page 208 for
details.
• No change to any VTAM RAS major node from the subarea environment.
Refer to D.2, “VTAM RAS Definitions” on page 233 for details.
Here we list only the changed major nodes. For the list of other major nodes on
RAA and RAS, refer to Appendix D, “VTAM Subarea Initial Definitions” on
page 207.
ATCSTR10
*******************************************************************
* *
* ATCSTR10 FOR RAA *
* THIS IS SUBAREA 10 STAGE 2: ICN *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @ICN1 ADD ICN FOR NN AND CONNTYPE=LEN *
* @ICN2 CHANGED DYNLU TO YES *
*******************************************************************
APPNCOS=#CONNECT, @ICN1* X
CDSERVR=YES, @ICN1* X
CONFIG=10, X
CONNTYPE=LEN, @ICN1* X
CPCP=YES, @ICN1* X
CSALIMIT=0, X
DIRSIZE=0, DEFAULT @ICN1* X
DIRTIME=8D, DEFAULT @ICN1* X
DYNADJCP=YES, DEFAULT @ICN1* X
DYNLU=YES, @ICN2* X
HOSTPU=ISTPUSA0, X
HOSTSA=10, X
HPR=NONE, @ICN1* X
INITDB=ALL, DEFAULT @ICN1* X
IOBUF=(100,384,5,,1,30), X
NETID=USIBMRA, X
NODETYPE=NN, @ICN1* X
NOTRACE,TYPE=VTAM,IOINT= X
NUMTREES=100, DEFAULT @ICN1* X
RESUSAGE=100, DEFAULT @ICN1* X
ROUTERES=128, DEFAULT @ICN1* X
SSCPID=99, X
SSCPNAME=RAA, SSCPNAME/CPNAME X
SORDER=SUBAREA, @ICN1* X
SSEARCH=YES, DEFAULT @ICN1* X
SUPP=NOSUP, X
TNSTAT,CNSL, X
XNETALS=YES
Once VTAM RAA has been migrated to an ICN, the data host RAS is migrated to
an MDH. The following is a summary of the changes required on RAA and RAS
for the implementation:
• On RAA:
− The configuration list ATCCON10 is updated.
− The CDRM major node RAACDRM is updated for VR-TG.
− New major nodes RAAAHHC and RAALCL2 are created.
− The MPC major node RAAMPC1 is no longer required.
− The following major nodes have not changed:
- ATCSTR10 (see F.1, “VTAM RAA Definitions Updated for ICN” on
page 244)
- RAAPATH, RAACDRSC, RAAADJ, all switched major nodes,
RA5NCPX and RA8NCPX (see D.2, “VTAM RAS Definitions” on
page 233)
• On RAS numerous major nodes are updated, so in the appendix here we list
all RAS′s major nodes.
ATCCON10
*******************************************************************
* ATCCON10 FOR RAA *
* THIS IS SUBAREA 10 AFTER MIGRATION OF SUBAREA 10 TO AN ICN *
* AND SUBAREA 28 TO AN MDH *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 REPLACE MPC WITH AHHC *
*******************************************************************
RAAATSO, TSO APPL X
RAAANETV, NETVIEW V2.4 APPL X
RAAPATH, SA10 PATH TABLE X
RAACDRM, CDRM X
RAACDRSC, CDRSCS X
RAACDILU, ILUS X
RAAADJ, ADJSSCP TABLE X
RAACTCA, CTC TO SA 20 X
RAAAHHC, ADD @MDH1 AHHC TO SA 28 (FID2) X
RAALCL1, LOCAL 3270 SCREENS X
RAALCL2, ADD @MDH1 AHHC TO SA 28 (FID2) X
RAASWMN1, SW-TR MN FOR 3174 X
RAASWMN2, SW-TR MN FOR CM/2 (KEN) X
RAASWMN3, SW-TR MN FOR CM/2 (ANAR) X
RAASWMN4, SW-TR MN FOR CM/2 (RABINDER) X
RAASWMN5, SW-TR MN FOR AS/400 X
RA5NCPX, NCP SA 5 X
RA8NCPX REMOTE NCP SA8
**RAAMPC1, DEL @MDH1 CTC TO SA 28 (FID4 MPC) X
RAACDRM
*******************************************************************
* CDRM MAJORNODE FOR RAA *
* SUBAREA 10 AS AN ICN AND SUBAREA 28 AS AN MDH *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 ADD PARAMETERS FOR VRTG LINK TO RAS *
*******************************************************************
VBUILD TYPE=CDRM
NETWORK NETID=USIBMRA
RAA CDRM SUBAREA=10,CDRDYN=YES,CDRSC=OPT
RAK CDRM SUBAREA=20,CDRDYN=YES,CDRSC=OPT
RAS CDRM SUBAREA=28,CDRDYN=YES,CDRSC=OPT, X
VRTG=YES @MDH1*
RAAAHHC
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAA (SA10) TO RAS (SA28) *
* WHEN RAA IS AN ICN AND RAS IS AN MDH *
****************************************************************
RAAAHHC VBUILD TYPE=TRL
* APPN MPC LINK FROM RAA TO RAS
RAAAHHCT TRLE LNCTL=MPC,READ=(C12),WRITE=(C14)
RAALCL2
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAA (SA10) TO RAS (SA28) *
* WHEN RAA IS AN ICN AND RAS IS AN MDH *
****************************************************************
RAALCL2 VBUILD TYPE=LOCAL
RAAPUAHC PU TRLE=RAAAHHCT, *
CONNTYPE=APPN, *
CPCP=YES, *
PUTYPE=2, *
XID=YES, *
TGP=CHANNEL, *
ISTATUS=ACTIVE
Appendix G. VTAM Definitions with RAA as ICN and RAS as MDH 247
G.2 VTAM RAS Definitions Updated for MDH
ATCSTR28
*******************************************************************
* ATCSTR28 FOR RAS *
* THIS IS SUBAREA 28 AFTER MIGRATION TO AN MDH NODE, AND *
* SUBAREA 28 IS AN ICN *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 ADD MDH FOR EN AND CONNTYPE=APPN *
*******************************************************************
CONFIG=28, X
MAXSUBA=255, X
CSALIMIT=0, X
CONNTYPE=APPN, DEFAULT @MDH1* X
CPCP=YES, @MDH1* X
DYNLU=YES, X
DYNADJCP=YES, DEFAULT @MDH1* X
GWSSCP=NO, X
HOSTPU=ISTPUS28, X
HOSTSA=28, X
HPR=NONE, @MDH1* X
IOBUF=(100,500,5,,1,30), X
NCPBUFSZ=2048, X
NETID=USIBMRA, X
NODETYPE=EN, @MDH1* X
NOTRACE,TYPE=VTAM,IOINT=0, X
NOTRACE,TYPE=VTAM,IOINT=0, X
PPOLOG=YES, X
SORDER=SUBAREA, @MDH1* X
SSCPID=28, X
SSCPNAME=RAS, X
SUPP=NOSUP, X
VRTG=NO, DEFAULT @MDH1 X
XNETALS=YES
ATCCON28
*******************************************************************
* ATCCON28 FOR RAS *
* THIS IS SUBAREA 28 AFTER MIGRATION OF SUBAREA 10 TO AN ICN *
* AND SUBAREA 28 TO AN MDH *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 REPLACE MPC WITH AHHC *
*******************************************************************
RASTSO, TSO APPL X
RASNETV, NETVIEW APPL X
RASCTCA, CTC TO SA28 X
RASLCL1, LOCAL 3270 DEFN X
RASLCL2, ADD @MDH1 AHHC TP SA10 (FID2) X
RASPATH, SA28 PATH TABLE X
RASCDRM, CDRM TABLE X
RASAHHC, ADD @MDH1 AHHC TO SA10 (FID2) X
RASADJ, ADJSSCP TABLE X
RASCA8 CA FOR NCP8
**RASMPC1, DEL @MDH1 MPC TO SA10 (FID4)
RASCDRM
*******************************************************************
* CDRM MAJORNODE FOR RAS *
* SUBAREA 10 AS AN ICN AND SUBAREA 28 AS AN MDH *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 ADD PARAMETERS FOR VRTG LINK TO RAA *
*******************************************************************
VBUILD TYPE=CDRM
NETWORK NETID=USIBMRA
RAS CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=28
RAA CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=10, X
VRTG=YES @MDH1*
RAK CDRM CDRDYN=YES,CDRSC=OPT,SUBAREA=20
Appendix G. VTAM Definitions with RAA as ICN and RAS as MDH 249
G.2.4 ANNC Definitions from RAS to RAA
RASAHHC
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAS (SA28) TO RAA (SA10) *
* WHEN RAA IS AN ICN AND RAS IS AN MDH *
****************************************************************
RASAHHC VBUILD TYPE=TRL
* APPN MPC LINK FROM RAS TO RAA
RASAHHCT TRLE LNCTL=MPC,READ=(C14),WRITE=(C12)
RASLCL2
****************************************************************
* AHHC DEFINITION FOR MPC LINK FROM RAS (SA28) TO RAA (SA10) *
****************************************************************
RASAHHC VBUILD TYPE=LOCAL
RASPUAHC PU TRLE=RASAHHCT, *
CONNTYPE=APPN, *
CPCP=YES, *
XID=YES, *
TGP=CHANNEL, *
ISTATUS=ACTIVE
With the conversion of VTAM RAS from an MDH to a pure EN, the VTAM
definitions are as follows:
• On RAA:
− Only NCP RA8NCPX has changed. The changes are listed below in H.1,
“VTAM RAA Definitions Updated for Pure EN” on page 252.
− No change to the remaining major nodes for RAA. They remain the
same as in Appendix G, “VTAM Definitions with RAA as ICN and RAS as
MDH” on page 245.
• On RAS:
− Major nodes ATCSTR28, ATCCON28, RASCA8 have changed from the
MDH definitions. These major nodes are listed below in H.2, “VTAM RAS
Definitions Updated for Pure EN” on page 253.
− The remaining major nodes on RAS are the same as the ones set up for
the MDH environment (see Appendix G, “VTAM Definitions with RAA as
ICN and RAS as MDH” on page 245).
Portion of RA8NCPX
*******************************************************************
* NCP DEFINITIONS FOR CHANNEL CONNECTION FROM RA8NCPX TO RAS *
* AFTER MIGRATION TO APPN WITH RAS AS EN (FID2 CONNECTION) *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @EN1 CHANGE FROM FID4 TO FID2 CONNECTION *
*******************************************************************
*
***********************************************************************
* CHANNEL ADAPTER LINKS #### *
***********************************************************************
RA8GCA GROUP LNCTL=CA, *
ISTATUS=INACTIVE
RA8CE20 LINE ADDRESS=0, *
CA=TYPE6-TPS, *
CASDL=120, *
CONNTYPE=APPN, ADD @EN1 *
DELAY=0.2, *
DYNADMP=NONE, *
ISTATUS=ACTIVE, ADD @EN1 *
NCPCA=ACTIVE, *
TIMEOUT=120 *
* STATOPT=(′ CA5 LINE ′ )
RA8PCA PU PUTYPE=2, UPD @EN1 *
XID=YES, FOR TYPE 2.1 FID2 ADD @EN1 *
TGP=CHANNEL FOR TYPE 2.1 FID2 ADD @EN1
* STATOPT=(′ CA5 PU ′ )
ATCSTR28
*******************************************************************
* ATCSTR28 FOR RAS *
* THIS IS SUBAREA 28 AFTER MIGRATION TO AN MDH NODE, AND *
* SUBAREA 28 IS AN ICN *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 ADD MDH FOR EN AND CONNTYPE=APPN *
* @EN1 DEL FOR PURE END NODE *
*******************************************************************
CONFIG=28, X
MAXSUBA=255, X
CSALIMIT=0, X
CONNTYPE=APPN, DEFAULT @MDH1* X
CPCP=YES, @MDH1* X
DYNLU=YES, X
DYNADJCP=YES, DEFAULT @MDH1* X
GWSSCP=NO, X
HOSTPU=ISTPUS28, X
HPR=NONE, @MDH1* X
IOBUF=(100,500,5,,1,30), X
NCPBUFSZ=2048, X
NETID=USIBMRA, X
NODETYPE=EN, @MDH1* X
NOTRACE,TYPE=VTAM,IOINT=0, X
PPOLOG=YES, X
SORDER=SUBAREA, @MDH1* X
SSCPID=28, X
SSCPNAME=RAS, X
SUPP=NOSUP, X
VRTG=NO, DEFAULT @MDH1 X
XNETALS=YES
**HOSTSA=28 DEL @EN1
ATCCON28
*******************************************************************
* ATCCON28 FOR RAS *
* THIS IS SUBAREA 28 AFTER MIGRATION OF SUBAREA 10 TO AN ICN *
* AND SUBAREA 28 TO AN MDH *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @MDH1 REPLACE MPC WITH AHHC *
* @EN1 ENTRIES NOT REQUIRED FOR PURE EN *
*******************************************************************
RASTSO, TSO APPL X
RASNETV, NETVIEW APPL X
RASCTCA, CTC TO SA28 X
RASLCL1, LOCAL 3270 DEFN X
RASLCL2, ADD @MDH1 AHHC TO SA10 (FID2) X
RASAHHC, ADD @MDH1 AHHC TO SA10 (FID2) X
RASADJ, ADJSSCP TABLE X
RASXCA, XCA FOR 3172 X
RASCA8 CA FOR NCP8
**RASMPC1, DEL @MDH1 MPC TO SA10 (FID4)
**RASPATH, DEL @EN1 SA28 PATH TABLE
**RASCDRM, DEL @EN1 SA28 CDRM TABLE
RASCA8
*******************************************************************
* CA MAJNODE FOR CHANNEL CONNECTION (TYPE2.1) TO RA8NCPX *
* AFTER MIGRATION TO APPN WITH RAS AS EN (FID2 CONNECTION) *
* *
* MARK CHANGE *
* ---- --------------------------------------- *
* @EN1 CHANGE FROM FID4 TO FID2 CONNECTION *
*******************************************************************
RASCA8 VBUILD TYPE=LOCAL
RASE20P PU PUTYPE=2, ADD @EN1* X
CUADDR=E20, ADD @EN1* X
XID=YES, ADD @EN1* X
CPCP=YES, ADD @EN1* X
SSCPFM=USSSCS, ADD @EN1* X
CONNTYPE=APPN ADD @EN1*
*RASE20G GROUP LNCTL=NCP DEL @EN1*
*RASE20L LINE ADDRESS=E20 DEL @EN1*
*RASE20P PU ISTATUS=ACTIVE,PUTYPE=4,MAXDATA=5000 DEL @EN1*
The NCP changes for the SSCP takeover were basic, and involved simply adding
BACKUP=YES and OWNER=DUMMY to the PCCU statement. Each GROUP
also had OWNER=DUMMY defined.
ATCSTR28
*******************************************************************
* ATCSTR28 FOR RAS *
* THIS IS SUBAREA 28 AS A BACKUP ICN FOR SSCP TAKEOVER *
* *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
APPNCOS=#CONNECT,
CONFIG=28,
BN=YES,
CDSERVR=YES,
CONNTYPE=APPN,
CONFIG=28,
CPCP=YES,
CSALIMIT=0,
DIRSIZE=0,
DIRTIME=8D,
DYNADJCP=YES,
DYNLU=YES,
GWSSCP=YES,
HOSTPU=ISTPUS28,
HOSTSA=28,
HPR=NONE,
INITDB=ALL,
IOBUF=(100,500,5,,1,30),
MAXSUBA=255,
NETID=USIBMRA,
NODETYPE=NN,
NOTRACE,TYPE=VTAM,IOINT=0,
NUMTREES=100,
RESUSAGE=100,
ROUTERES=128,
PPOLOG=YES,
SORDER=SUBAREA,
SSCPID=10028,
SSCPNAME=RAS,
SSEARCH=YES,
SUPP=NOSUP,
TNSTAT,CNSL,
VRTG=NO,
XNETALS=YES
RASCA8
RASCA8 VBUILD TYPE=CA
* CHANNEL LINK FROM 28 TO 8
RASE20G GROUP LNCTL=NCP
RASE20L LINE ADDRESS=E20
RASE20P PU ISTATUS=ACTIVE,PUTYPE=4,MAXDATA=5000
RAAADJCP
*******************************************************************
* *
* ADJCP TABLE FOR RAA *
* *
* SG24-4656 SUBAREA TO APPN MIGRATION GUIDELINES *
* *
*******************************************************************
VBUILD TYPE=ADJCP
CP05137 ADJCP NETID=USIBMRA,DYNLU=YES
CP05147 ADJCP NETID=USIBMRA,DYNLU=YES
CP05160 ADJCP NETID=USIBMRA,DYNLU=YES
CP31742 ADJCP NETID=USIBMRA,DYNLU=YES
CP31744 ADJCP NETID=USIBMRA,DYNLU=YES
This appendix describes the exit ISTEXCCS, supplied with VTAM to allow
dependent LUs and their PUs to be defined dynamically. If VTAM finds this exit
in your VTAMLIB data set it will be used.
Character(s) Convention
A - M, @, $, # Used for customer generated names.
(IBM generated names will not begin
with these characters).
N Used for dynamic I/O generated names.
(customer seed not supplied)
O - T Reserved for IBM use.
U - Z Used to identify IBM generated names.
Note:
If you do not wish to update your exit, you can use two definition files, CPNDEF
and NIDDEF. These are 80-byte fixed record sequential data sets which are
allocated by VTAM′s start procedure. The configuration services XID exit looks
for these data sets when it is invoked.
Using these files gives you control over the names you want to define in your
network. At the same time it allows you to point to the model major node entry
of your choice for a given device. If the CP name or the IDBLK/IDNUM are not
found in these data sets the exit will attempt to build a name from the name
generation table. In either case the definitions are added to ISTDSWMN. If the
As seen previously in Figure 137 on page 258, definitions are added to the
dynamically defined major node ISTDSWMN, created by VTAM. You have control
over the names defined in VTAM by this method. The exit must be used in
conjunction with a model major node.
Independent LUs are not supported in this exit. Use predefined CDRSCs or
allow dynamic ILU definition.
New VTAM functions are continually being added by means of APARs rather
than in new releases. This has the benefit of making customer-requested
features available earlier than they would otherwise appear. However, the
disadvantage is that there is no formal method of broadcasting the changes.
Therefore, we have listed here the most recent APPN-related functions that have
been introduced in this way. Our base is VTAM V4R4, but some of these may
have been retro-fitted to earlier releases.
Reference to PTF numbers that have not been released through the normal
distribution process does not imply general availability. The purpose of
including these reference numbers is to alert IBM customers to specific
information relative to the implementation of the PTF when it becomes available
to each customer according to the normal IBM PTF distribution process.
Information in this book was developed in conjunction with use of the equipment
specified, and is limited in application to those specific hardware and software
products and levels.
IBM may have patents or pending patent applications covering subject matter in
this document. The furnishing of this document does not give you any license to
these patents. You can send license inquiries, in writing, to the IBM Director of
Licensing, IBM Corporation, 500 Columbus Avenue, Thornwood, NY 10594 USA.
Licensees of this program who wish to have information about it for the purpose
of enabling: (i) the exchange of information between independently created
programs and other programs (including this one) and (ii) the mutual use of the
information which has been exchanged, should contact IBM Corporation, Dept.
600A, Mail Drop 1329, Somers, NY 10589 USA.
The information contained in this document has not been submitted to any
formal IBM test and is distributed AS IS. The information about non-IBM
(″vendor″) products in this manual has been supplied by the vendor and IBM
assumes no responsibility for its accuracy or completeness. The use of this
information or the implementation of any of these techniques is a customer
responsibility and depends on the customer′s ability to evaluate and integrate
them into the customer′s operational environment. While each item may have
been reviewed by IBM for accuracy in a specific situation, there is no guarantee
that the same or similar results will be obtained elsewhere. Customers
attempting to adapt these techniques to their own environments do so at their
own risk.
Reference to PTF numbers that have not been released through the normal
distribution process does not imply general availability. The purpose of
including these reference numbers is to alert IBM customers to specific
information relative to the implementation of the PTF when it becomes available
to each customer according to the normal IBM PTF distribution process.
Microsoft, Windows, Windows NT, and the Windows 95 logo are trademarks
or registered trademarks of Microsoft Corporation.
The publications listed in this section are considered particularly suitable for a
more detailed discussion of the topics covered in this redbook.
This information was current at the time of publication, but is continually subject to change. The latest
information may be found at http://www.redbooks.ibm.com/.
Redpieces
For information so current it is still in the process of being written, look at ″Redpieces″ on the Redbooks Web
Site ( http://www.redbooks.ibm.com/redpieces.html). Redpieces are redbooks in progress; not all redbooks
become redpieces, and sometimes just a few chapters will be published this way. The intent is to get the
information out much quicker than the formal publishing process allows.
IBMMAIL Internet
In United States: usib6fpl at ibmmail usib6fpl@ibmmail.com
In Canada: caibmbkz at ibmmail lmannix@vnet.ibm.com
Outside North America: dkibmbsh at ibmmail bookshop@dk.ibm.com
• Telephone orders
Redpieces
For information so current it is still in the process of being written, look at ″Redpieces″ on the Redbooks Web
Site ( http://www.redbooks.ibm.com/redpieces.html). Redpieces are redbooks in progress; not all redbooks
become redpieces, and sometimes just a few chapters will be published this way. The intent is to get the
information out much quicker than the formal publishing process allows.
Company
Address
We accept American Express, Diners, Eurocard, Master Card, and Visa. Payment by credit card not
available in all countries. Signature mandatory for credit card payment.
F
FID2 11 M
FID4 11 MDH 5
FID5 11 migration data host 5, 22, 26, 53, 95, 182
FQPCID 16 mode table entry 38
mode to COS resolution 167, 171, 178
multilink TGs 7
G multinode persistent sessions 181
GDS 17 multipath channel 39, 40
T
P TCID 16
Parallel Sysplex 181 TG numbers 107
parallel TGs 7 topology
peripheral border node 6 See Advanced Peer-to-Peer Networking topology
topology database 13, 33
transmission group 7, 14
R transmission group characteristics
rapid transport protocol 4, 10, 16
See Advanced Peer-to-Peer Networking TG
REGISTER keyword 99
characteristics
route selection control vector 17
transport connection identifier 16
route setup 18
transport resource list 40
RSCV 17
TRL major node 40
RSCV pruning 139
type 2 node 8
type 2.1 node 3, 4, 7, 8, 43
type 4 node 3
S type 5 node 3
SAMAP VTAMLST member 163
SATOAPPN COS mapping table 172
SATOAPPN VTAMLST member 38
searching
V
virtual route-based transmission group 8, 96, 114,
See Advanced Peer-to-Peer Networking searching
128, 131, 139
See subarea searching
VTAM
self-defining dependent LU exit 142
APPN connection to NCP 109
session services extensions 16, 28, 48, 99, 117, 171,
APPN COS table 64
173
APPN-related commands 203
SNI 50, 128, 138
COS mapping tables 38, 172, 179
SSCP 3, 8
CP-CP session priority 65
SSCP takeover 29, 141, 160
data sets 64
SSCP-SSCP sessions 8
node types in a mixed APPN/subarea network 195
start options
resource represntation 153
See VTAM start options
routing in mixed APPN/subarea networks 161
subarea matching 162
subarea COS for CP-CP sessions 178
subarea/APPN routing decision 151
Index 275
VTAM (continued) VTAM displays (continued)
TG profiles 64, 134 APPN topology 67, 68, 79, 82, 84, 134
topology agent 193 APPN topology with VR-TG 137
topology and directory checkpointing 64 APPN-related start options 66, 102
using DISPLAY commands in APPN networks 191 BN-related start options 123
VARY command in APPN networks 192 CP name invalid 31
VTAM definitions dynamic CDRSC definition failure 45
adjacent cluster table 52, 120 dynamic CDRSC major node 91
adjacent CP major node 45, 256 EN initialization 110
adjacent SSCP table 54, 235 LEN link station 68, 69, 83
adjacent SSCP tables 210 MDH initialization 101
ANNC link station 98, 247, 250 NN servers for EN 105
APPN channel link station to NCP 254 subarea routes 101, 104
APPN link station for network border 122 topology for border node 123, 125
APPN link station for same=NetID border 122 TSO as remote APPN LU 106
CA major node for APPN connection 109 VARY REL for SSCP takeover 147
CDRM major node 210, 235, 247, 249 VR-TG activation 137
CDRM major node with VR-TG 133 VTAM end node from itself 111
CDRSC major node 210 VTAM start options
configuration list 208, 233, 246, 249, 254 APPN-related options 197
COS mapping tables 172 APPNCOS 36, 63, 179
local non-SNA 3174 234 BN 50, 119
local SNA 3174 211 BNDYN 51, 119, 129
mode table 239 BNORD 51, 119, 161
mode table with APPN COS entries 172 CDRSCTI 55
model major node for configuration services CDSERVR 46, 62
exit 241 CDSREFER 46
multipath channel connections 40 CONNTYPE 43, 62, 65, 69, 95
NCP 215, 224, 252 converting CMC to ICN 62
network node server list 48, 99 converting data host to MDH 95
path tables 209, 236 CPCP 43, 63, 96
SDLC-attached 3174 71 DIRSIZE 46, 63
SDLC-attached CS/2 gateway 87 DIRTIME 46
start options 208, 233, 244, 248, 253, 255 DYNADJCP 44, 63, 96
subarea channel link to NCP 256 DYNDLGMD 167, 177, 178
subarea connection to NCP 235 DYNLU 44
subarea mapping table 163 DYNMODTB 167, 177, 178
subarea MPC connection 211, 234 for SSCP takeover 145
switched APPN connection 76, 83 HOSTSA 63, 96, 109
switched LEN connection 81 HPR 63, 96
switched major node for 3174 212 HPRNCPBF 162, 185
switched major node for AS/400 214 INITDB 46, 64
switched major node for CM/2 213 ISTCOSDF 36, 171, 179
TG profiles 37 NETID 51
transmission group profiles 237 NODETYPE 63, 96
transport resource list 40, 98 SNVC 51, 119, 129
transport resource list for ANNC 250 SORDER 54, 63, 96, 113, 156, 160
transport resource list for MPC 247 SRCHRED 56
VTAM displays SRCOUNT 56
active MPC TRLE 103 SRTIMER 56
adjacent APPN CP 72, 80, 85 SSCPNAME 63
adjacent APPN NN 91 SSCPORD 55, 152, 156, 161
adjacent cluster table 124, 126, 127 SSEARCH 55, 156, 159, 161
adjacent SSCP 100 summary of APPN-related options 56
adjacent SSCP table 54 VFYRED 154
adjacent VTAM CP and SSCP 103 VRTG 96, 133
ANNC activation 102 VRTGCPCP 133, 139
APPN link station 32, 45, 73, 90 XNETALS 51, 122
Your feedback is very important to help us maintain the quality of ITSO redbooks. Please complete this
questionnaire and return it using one of the following methods:
• Use the online evaluation form found at http://www.redbooks.ibm.com
• Fax this form to: USA International Access Code + 1 914 432 8264
• Send your comments in an Internet note to redbook@us.ibm.com
Please rate your overall satisfaction with this book using the scale:
(1 = very good, 2 = good, 3 = average, 4 = poor, 5 = very poor)
Overall Satisfaction ____________
Was this redbook published in time for your needs? Yes____ No____
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
_____________________________________________________________________________________________________
IBML
SG24-4656-01