Acme SBC Config

Avaya Solution & Interoperability Test Lab
Sample Configuration Illustrating Avaya Aura

Communication Manager SIP Trunking Using Processor
Ethernet and Acme Packet Net-Net 4500 Session Director
Issue 1.0
Abstract
These Application Notes illustrate a sample configuration of Avaya Aura Communication
Manager Release 5.2 with SIP Trunks to Acme Packet Net-Net 4500 Session Director at two
sites. For business continuity, a primary site uses Avaya S8730 Servers, and a secondary site
uses an Avaya S8500 Server as an Enterprise Survivable Server (ESS). At each site, two
Avaya C-LAN cards as well as the Processor Ethernet of the sites Avaya Server are
configured for SIP Trunking to the inside realm of an Acme Packet Net-Net 4500 Session
Director. On the outside realm, each Acme Packet Net-Net 4500 Session Director is
connected to a SIP network simulating a public SIP Service Provider. Within each site, the
Acme Packet Net-Net 4500 Session Director is configured for load spreading and fast fail-over
of inbound calls to the enterprise from the PSTN. For outbound calls to the PSTN,
Communication Manager is configured for location-based routing for trunk selection and
Look-Ahead Routing for trunk fail-over.
These Application Notes complement previously published Application Notes, focusing on the
new Communication Manager 5.2 capability to use the Processor Ethernet of the Avaya
S8730 Servers and Avaya ESS as the Avaya interface for SIP Trunk signaling.
JRR; Reviewed:
SPOC 7/21/2009
Solution & Interoperability Test Lab Application Notes

2009Avaya Inc. All Rights Reserved.
1 of 70
CM-PE-NN4500
1. Introduction
These Application Notes illustrate a sample configuration of Avaya Aura Communication
Manager Release 5.2 with SIP Trunks to Acme Packet Net-Net 4500 Session Director at two
sites. A primary site uses Avaya S8730 Servers, and a secondary site uses an Avaya S8500
Server as an Enterprise Survivable Server (ESS). At each site, two Avaya C-LAN cards as well
as the Processor Ethernet of the sites Avaya server are configured for SIP Trunking to the
Session Director using TCP for the SIP signaling connectivity. Each Session Director is also
connected to a SIP network simulating a SIP Service Provider. Since most SIP Service Providers
use UDP for SIP signaling, the SIP signaling connectivity from the Acme Packet Net-Net 4500
toward the outside realm uses UDP.
These Application Notes complement previously published Application Notes [JRR] that
covered the C-LAN based SIP Trunking and survivability considerations when various failures
are induced. Figure 1 is repeated from reference [JRR], but the detailed configuration and call
verifications from reference [JRR] will not be repeated in these Application Notes. These
Application Notes build upon the configuration documented in [JRR] by adding SIP Trunking
from the Processor Ethernet of the Avaya S8730 Servers, as well as the Avaya S8500 ESS.
Figure 1: Avaya Aura Communication Manager Survivable SIP Trunking
JRR; Reviewed:
SPOC 7/21/2009

2 of 70
CM-PE-NN4500
The network shown in Figure 1 remains in place for the verification of these Application Notes.
Figure 2 omits many of the details, but illustrates (in orange) the addition of two new SIP trunks
(32, 62) via the Processor Ethernet of the Avaya S8730 Servers and Avaya S8500 ESS.
Figure 2: Avaya Aura Communication Manager Processor Ethernet SIP Trunking

Although a mix of C-LAN based SIP trunks and Processor Ethernet based SIP trunks are used in
the sample configuration, the use of the Avaya S8730 Server Processor Ethernet does not require
Avaya G650 Media Gateways. In some configurations, Processor Ethernet can obviate the need
for Avaya G650 Media Gateways and the associated IPSI, C-LAN, and Media Processor cards.
Although not the subject of these Application Notes, Avaya H.248 gateways (e.g., Avaya G250,
G350, G450, or G430 Media Gateways) and Avaya H.323 Telephones can register directly to the
Avaya S8730 Processor Ethernet, beginning in Communication Manager 5.2. A mix of C-LAN
based SIP trunks and Processor Ethernet based SIP trunks are used in the sample configuration,
since it is expected that large customers interested in deploying a multi-site survivable SIP
Trunking solution may either already have Avaya G650 Media Gateways or require Avaya G650
Media Gateways for scale or specific application reasons. Combining the more familiar C-LAN
based SIP trunks with new Processor Ethernet based SIP trunks in the same configuration also
provides an opportunity to compare and contrast behaviors for instructional purposes.
The Acme Packet Net-Net 4500 is used to distribute SIP signaling for incoming calls to multiple
C-LAN interfaces and the Processor Ethernet of the sites Avaya server, providing load
spreading and fast automatic fail-over. The Acme Packet Net-Net 4500 performs conversion
between TCP transport for SIP signaling used by Communication Manager to UDP transport
commonly used by SIP Service Providers. The Acme Packet Net-Net 4500 also performs
Session Border Controller (SBC) functions, providing security and topology-hiding at the
enterprise edge. In the sample configuration, all SIP signaling and RTP media between the
JRR; Reviewed:
SPOC 7/21/2009

3 of 70
CM-PE-NN4500
enterprise and the (simulated) SIP Service Provider flows through the Acme Packet Net-Net
4500.
A customer interested in SIP Trunk survivability may want a redundant pair of Acme Packet
Net-Net 4500 Session Directors at each site. Although the sample configuration verified in these
Application Notes used only a single Acme Packet Net-Net 4500 at each site, the Acme Packet
configuration shown in Appendix A of reference [JRR] was prepared as if there were a high
availability Acme Packet configuration at each site. Actual verification testing of the Acme
Packet Net-Net 4500 High Availability configuration with Communication Manager was
performed as part of Avaya DevConnect compliance testing, and the Application Notes in
reference [AP-HA] documents the configuration and testing results.
1.1. Summary of Inbound Calls to the Enterprise

Figure 2 is repeated from reference [JRR] to illustrate aspects of the sample configuration
related to inbound calls to the enterprise from the PSTN. Although further elaboration of the
simulation of the SIP Service Provider is out of scope, Figure 2 may help with understanding
assumptions and the call flow verifications in Section 5. For example, it is assumed that
published PSTN telephone numbers such as Direct Inward Dial (DID), Listed Directory
Numbers (LDN), or toll-free numbers that map to Avaya Vector Directory Numbers (VDNs) can
arrive to the enterprise from the service provider via either the primary site or the secondary site
SIP Trunks. A SIP Service Provider may load balance inbound calls to the enterprise between
sites, or route particular numbers to a specific site preferentially, with fail-over to the other site
as needed.
JRR; Reviewed:
SPOC 7/21/2009

4 of 70
CM-PE-NN4500
As shown in Figure 3, the DID number 732-852-1816 is preferentially routed to the primary site,
but can fail-over to the secondary site. The DID number 732-852-2940 is preferentially routed to
the secondary site, but can fail-over to the primary site. Communication Manager can map any
received telephone number to any destination via the incoming call handling table of the trunk
group receiving the call. During testing, calls arriving via the primary site SIP trunks were
directed to users at both the primary and the secondary sites, and vice-versa.
Figure 3: Incoming Calls to the Enterprise from the PSTN
JRR; Reviewed:
SPOC 7/21/2009

5 of 70
CM-PE-NN4500
1.2. Session Director and Communication Manager Terminology

The table below is repeated from reference [JRR] to provide a translation for key concepts and
terminology that may be helpful to readers familiar with Acme Packet Session Director or Avaya
Aura Communication Manager, but not both. Of course, these analogies are imperfect, so the
table is intended only as a starting point in understanding.
Acme Packet
Concept
Session-agent
(SIP)
Session-agent
group (SAG)
Sag-recursion
Session-agent
configuration for
ping-method
OPTIONS
Avaya Aura
Communication
Manager Concept
SIP Signaling Group and
SIP Trunk Group
Route-Pattern with
multiple SIP Trunk
Groups
Notes
The session-agent defines a SIP peer, similar to

an Avaya SIP signaling group/trunk group.
Session agents can be configured in a SAG, so
that routing to the SAG can use any session
agent in the group. The analogy is imperfect in
that the SAG can use strategies that select agents
from the SAG based on real-time traffic
conditions and defined constraints (i.e., more
sophisticated outbound load balancing options)
Look-Ahead Routing
If a session agent is selected, but a
(LAR) for a routedownstream failure results, sag-recursion can
pattern
trigger Session Director to automatically use a
different session agent in the SAG. This is
similar to Communication Manager selection of
a SIP trunk group, followed by a downstream
signaling failure causing LAR (route-advance)
to the next trunk in the route pattern.
Enable Layer 3 Test =
Both Communication Manager and Acme
y for the SIP Signaling Packet can be configured to source OPTIONS
Group
messages to check the health of a peer. See
Section 1.3 for more information.
To summarize the Acme Packet Net-Net 4500 configuration, a session agent is defined for
each C-LAN at the site, and the processor Ethernet of the Avaya S8730 Server or Avaya
S8500 ESS, depending on the site. A session agent group (SAG) is configured to group the
session agents, and a strategy for distribution of calls to the session agents that are members of
the group is specified. In the sample configuration, once an inbound call reaches an enterprise
site, the Session Director is configured for round-robin call distribution to the members of the
session agent group. If connectivity to a particular session agent fails, and the Acme Packet NetNet 4500 has not yet detected the failure, an inbound call directed to the failed session agent will
encounter a transaction timeout. The transaction timeout will cause the call to be directed to
another session agent at the same site automatically. In addition, the failed session agent will be
marked out-of-service so that subsequent inbound calls will flow to an operational session agent
without experiencing the timeout. If all session agents that are part of the session agent group
experience a transaction timeout, then a SIP 408 message would be returned to the SIP Service
JRR; Reviewed:
SPOC 7/21/2009

6 of 70
CM-PE-NN4500
Provider. If all session agents that are part of the session agent group are already marked out of
service, then a SIP 503 Service Unavailable would be returned to the SIP Service Provider.
Similarly, if the public side of the Session Director experienced failures for an outbound call
from Communication Manager, Communication Manager would receive no SIP response after
100 TRYING, a 408 Transaction Timeout, or a 503 Service Unavailable, depending on the
particular failure scenario. Note that all these conditions are triggers for Communication
Manager Look-Ahead Routing. Reference [LAR] documents another sample configuration for
Look-Ahead Routing, and includes a more complete list of SIP triggers.
In the sample configuration, the Acme Packet Net-Net 4500 at a given site does not have session
agents to the other site. It is presumed that a production SIP Service Provider can redirect calls
from one site to another based on failure conditions, such as a timeout, or the return of a 503
Service Unavailable.
1.3. SIP OPTIONS Message, Service States, and Call Acceptance

Both Communication Manager and the Acme Packet Net-Net 4500 can use a SIP OPTIONS
message to verify connectivity health. This section summarizes the use of the SIP OPTIONS
message, the implications for in-service and out-of-service determinations, and the effect on new
call attempts. See Section 5.7 of reference [JRR] as well as Section 5.7 of these Application
Notes for Wireshark traces related to the topics in this section.
In the sample configuration, the Acme Packet Net-Net 4500 is configured to periodically check
the availability of a session agent via a SIP OPTIONS message. The interval between SIP
OPTIONS messages is configurable. By default, any SIP response would be considered an
acceptable reply, including normal 200 OK responses, but also other responses such as 503
Unavailable. The responses from Communication Manager deemed acceptable for marking the
session agent in-service can be configured, if desired. Although a failed SIP OPTIONS
exchange can result in a session agent being marked out-of-service, in a system with continuous
call activity, it would be more likely that a transaction timeout for a SIP method such as INVITE
would cause a recently failed session agent to be marked out-of-service. In this light, the SIP
OPTIONS exchange is more likely to be the method of bringing a previously failed session-agent
back in service. Therefore, if rapid recovery from prior failures is paramount, the time between
SIP OPTIONS generated by the Acme Packet Net-Net 4500 can be reduced to a low value. In
the sample configuration, testing was done with a 60 second interval, and later with a 16 second
interval.
If the Acme Packet Net-Net 4500 has marked a session-agent out-of-service, the session agent
will not be chosen for call activity. That is, if both an out-of-service session agent and an inservice session agent appear in the same session agent group, the in-service session agent will
naturally be chosen for the next call. The out-of-service session agent can come back in-service
either via a SIP response to an Acme Packet sourced SIP OPTIONS message, or due to a SIP
message, such as an INVITE or OPTIONS received from the session agent. Indeed, if an
INVITE message for a Communication Manager outbound call is received by the Acme Packet
Net-Net 4500 from a session agent that had been marked out-of-service, the INVITE is
processed, the call can succeed, and the session agent is again marked in-service. For the reader
JRR; Reviewed:
SPOC 7/21/2009

7 of 70
CM-PE-NN4500
familiar with Avaya trunk states, this is similar to the Communication Manager behavior for a
trunk marked in the Out-of-service/Far-end state.
When an Avaya SIP signaling group is marked with Enable Layer 3 Test = y,
Communication Manager will periodically send a SIP OPTIONS method to the far-end of the
signaling group. When the Acme Packet Net-Net 4500 receives such a SIP OPTIONS, it checks
the logical next-hop. In the sample configuration, the next hop is the SIP Service Provider.
If there is no in-service next-hop, then the Acme Packet Net-Net 4500 returns a 503 Service
Unavailable to Communication Manager. Communication Manager will then mark the SIP
signaling group for bypass, and the corresponding SIP trunk group will be marked Out-ofservice/Far-end. For example, if the Acme Packet Net-Net 4500 has detected that the SIP
Service Provider network has failed, then a SIP OPTIONS from the Avaya session agent will
receive a 503, and the Avaya trunks will be marked for bypass, which is appropriate. In this
state, although outbound calls from the enterprise will not select the trunk, if an inbound call is
received, the network has apparently recovered. The call will be accepted, and the Avaya SIP
trunk group will be marked in-service.
1.4. Summary of Outbound Calls from the Enterprise to the PSTN

For outbound calls, Communication Manager location-based routing can direct outgoing calls
from users at a given site to the SIP trunks in the same site, with fail-over to use SIP trunks at the
other site as needed. The user dials the Automatic Route Selection (ARS) access code followed
by the PSTN number. In the sample configuration, if a user at the primary site dials the number,
the call will be directed to route-pattern 30, which lists the SIP trunks at the primary site first. If
the trunks at the primary site are unable to take the call, either due to congestion or failure, the
call will proceed out the trunks at the secondary site, which are also members of route-pattern 30.
Outbound calls placed from users at the secondary site are directed to route-pattern 60, which
lists the SIP trunks at the secondary site first, with overflow and fail-over to the SIP trunks at the
primary site.
JRR; Reviewed:
SPOC 7/21/2009

8 of 70
CM-PE-NN4500
2. Equipment and Software Validated

The following equipment and software were used for the sample configuration provided:
Equipment
Avaya 4600-Series Telephones (H.323)
Software
Avaya Aura Communication Manager
Release 5.2 (947.3) with and without SP1
(patch 17294)
Avaya Aura Communication Manager
Release 5.2 (947.3) with and without SP1
(patch 17294)
Release 2.9 H.323
Avaya 9600-Series Telephones (H.323)
Release 3.0 H.323
Avaya 2400-Series Digital Telephones

Acme Packet Net-Net 4500 Session Directors
Cisco AS5400 Universal Gateway
N/A
CX6.1.0 patch 3
12.4(15)T7
Cisco 3825 Integrated Services Router
12.4(11)XW7
Avaya S8730 Servers (Main)
Avaya S8500 Server (ESS)
Table 1: Equipment and Software Versions Used
3. Avaya Aura Communication Manager Configuration

The network shown in Figure 1 is documented in reference [JRR]. Rather than repeat the
configuration, this section focuses on the changes to the configuration to allow SIP signaling via
the Processor Ethernet of the Avaya S8730 Servers and Avaya S8500 ESS. Product
documentation can be found in references [CM1], [CM2], [CM3], and [ESS]. Except for the
web page configuration shown in Section 3.1, all remaining configuration is performed via the
Communication Manager SAT interface. Screens are abridged for brevity in presentation.
3.1. Configuration Processor Ethernet Via the Web Pages

The Processor Ethernet must be configured via the Configure Server Web pages on both
S8730 servers. The screens in this section illustrate the configuration on only one of the servers.
Consult product documentation for further procedural guidance (e.g., performing changes on the
standby server to mitigate service disruptions).
JRR; Reviewed:
SPOC 7/21/2009

9 of 70
CM-PE-NN4500
The S8730 Server can be accessed via a web interface in an internet browser. The following
screen shows the initial screen. To add Processor Ethernet to an existing configuration such as
the configuration documented in [JRR], select Configure Server under Installation, as shown.
Navigate to Set Identities. The following screen shows the lower portion of the Set Identities
page before any changes are made. Note that the Processor Ethernet (PE) drop-down initially
indicates UNUSED.
JRR; Reviewed:
SPOC 7/21/2009

10 of 70
CM-PE-NN4500
The following screen shows the lower portion of this same screen, where Ethernet 2 is being
selected in the Processor Ethernet (PE) drop-down. In this case, Ethernet 2 is the Corporate
LAN interface, and the Processor Ethernet and Corporate LAN interface will be the same.
Select the Continue button to proceed. The screen below shows a portion of the resulting
screen, with the IP Addresses from the sample configuration populated in the mandatory fields,
indicated by a red star. The active server IP Address is 2.2.87.13. This is the IP Address that
will be associated with the Processor Ethernet or procr of the active S8730 Server.
JRR; Reviewed:
SPOC 7/21/2009

11 of 70
CM-PE-NN4500
At the conclusion of the process, a screen such as the following should be displayed.
At the conclusion of the process on both S8730 Servers, the Status Summary page may be used
to check the Processor Ethernet status, as shown below.
JRR; Reviewed:
SPOC 7/21/2009

12 of 70
CM-PE-NN4500
3.2. Node Names

Node names are mappings of names to IP Addresses that can be used in various screens. The
following abridged list output shows some of the relevant node-names in the sample
configuration. The entry for the Processor Ethernet called procr is shown in bold. Note that
the procr node-name appears automatically due to the web configuration in Section 3.1.
list node-names all
Page
NODE NAMES
Type
IP
IP
IP
IP
IP
IP
IP
IP
Name
nn4500-prisite
nn4500-secsite
procr
tn2602-1a11
tn2602-2a07
tn2602-2b07
val-1a07
val-2a08
IP Address
2.2.85.45
2.2.185.145
2.2.87.13
2.2.185.4
2.2.26.3
2.2.26.2
2.2.185.25
2.2.85.25
3.3. Adding the IP Interface for procr

The add ip-interface procr command is used to configure the Processor Ethernet parameters.
The following screen shows the parameters used in the sample configuration. Although not the
focus of these Application Notes, it can be observed that the Processor Ethernet may also be used
for registrations from H.323 IP Telephones and H.248 gateways.
add ip-interface procr
Page
1 of
IP INTERFACES
Type: PROCR
Target socket load: 19200
Enable Interface? y
Allow H.323 Endpoints? y

Allow H.248 Gateways? y
Gatekeeper Priority: 5
Network Region: 3
IPV4 PARAMETERS
Node Name: procr
Subnet Mask: /24
JRR; Reviewed:
SPOC 7/21/2009

13 of 70
CM-PE-NN4500
The following screen lists the IP interfaces, which now includes the bold procr at the end of
the list.
list ip-interface all
Page
IP INTERFACES
ON Type
Slot
Code/Sfx
-- ------ ----- -------y C-LAN 01A02 TN799 D

y C-LAN
02A02 TN799
y VAL
01A07 TN2501
y MEDPRO 02A07 TN2602

y MEDPRO 02B07 TN2602
y MEDPRO 01A11 TN2602
y C-LAN
02B02 TN799
y C-LAN
01A10 TN799
y VAL
02A08 TN2501
y PROCR
Node Name/
IP-Address
--------------c-lan
2.2.185.2
c-lan2a02
2.2.85.2
val-1a07
2.2.185.25
tn2602-2a07
2.2.26.3
tn2602-2b07
2.2.26.2
tn2602-1a11
2.2.185.4
c-lan2b02
2.2.85.20
c-lan1A10
2.2.185.20
val-2a08
2.2.85.25
2.2.87.13
Mask
Gateway Node
Net
Rgn
VLAN
---/24
--------------Gateway001
--1
---n
/24
Gateway002
/24
Gateway001
/24
Gateway003
/24
Gateway003
/24
Gateway001
/24
Gateway002
/24
Gateway001
/24
Gateway002
/24
2.2.87.1
n
3
3.4. SIP Signaling Groups Using Processor Ethernet

This section illustrates the configuration of the SIP Signaling Groups that use the Processor
Ethernet. The SIP Signaling Groups that use a C-LAN are documented in reference [JRR]. Two
additional signaling groups are defined. Each signaling group has a Group Type of sip, and
a Near-end Node Name of procr. The Transport Method for all signaling groups is tcp
using port 5060. The Far-end Domain for each signaling group is the inside IP Address of
the appropriate Acme Packet Net-Net 4500. The Enable Layer 3 Test field is enabled to allow
Communication Manager to maintain the signaling group using the SIP OPTIONS method, as
described in Section 1.3. Other fields can be left at default values, including DTMF over IP
set to rtp-payload, which corresponds to RFC 2833. Note that the Alternate Route Timer
that defaults to 6 seconds impacts fail-over timing for outbound calls. If Communication
Manager does not get an expected response, Look-Ahead Routing can be triggered, after the
expiration of the Alternate Route Timer.
JRR; Reviewed:
SPOC 7/21/2009

14 of 70
CM-PE-NN4500
The following screen shows the addition of signaling group 32. The Near-end Node Name is
procr and the Far-end Node Name is the node name of the Acme Packet Net-Net 4500 at the
primary site. The Far-end Network Region is configured to region 3. In normal operation,
when all network elements are up, signaling group 32 will be in-service, under the control of the
active S8730 Server. If the ESS becomes active, signaling group 32 will be out-of-service from
the point of view of the ESS. The ESS will never receive a response from the primary site
Session Director at the far-end of signaling group 32.
add signaling-group 32
Group Number: 32
Page
1 of
Group Type: sip

Transport Method: tcp
IMS Enabled? n
Near-end Node Name: procr
Near-end Listen Port: 5060
Far-end Node Name: nn4500-prisite

Far-end Listen Port: 5060
Far-end Network Region: 3
Far-end Domain: 2.2.85.45

Bypass If IP Threshold Exceeded? n
DTMF over IP: rtp-payload
Session Establishment Timer(min): 3
Enable Layer 3 Test? y
H.323 Station Outgoing Direct Media? n
Direct IP-IP Audio Connections?

IP Audio Hairpinning?
Direct IP-IP Early Media?
Alternate Route Timer(sec):
y
n
n
6
The following screen shows the addition of signaling group 62. The Near-end Node Name is
procr and the Far-end Node Name is the node name of the Acme Packet Net-Net 4500 at the
secondary site. The Far-end Network Region is configured to region 1. In normal operation,
when all network elements are up, signaling group 62 will be out-of-service from the point of
view of the active S8730 Server, which will never receive a SIP 200 OK response from the
secondary site Session Director at the far-end of signaling group 62. However, when the ESS
is active, signaling group 62 will come in-service from the point of view of the ESS.
add signaling-group 62
Group Number: 62
Page
1 of
Group Type: sip

Transport Method: tcp
IMS Enabled? n
Near-end Node Name: procr
Near-end Listen Port: 5060
Far-end Node Name: nn4500-secsite

Far-end Listen Port: 5060
Far-end Network Region: 1
Far-end Domain: 2.2.185.145

Bypass If IP Threshold Exceeded? n
DTMF over IP: rtp-payload
Session Establishment Timer(min): 3
Enable Layer 3 Test? y
H.323 Station Outgoing Direct Media? n
Direct IP-IP Audio Connections?

IP Audio Hairpinning?
Direct IP-IP Early Media?
Alternate Route Timer(sec):
y
n
n
6
3.5. SIP Trunk Groups

This section illustrates the configuration of two additional SIP Trunks Groups, corresponding to
the two SIP signaling groups using Processor Ethernet. Each trunk group has a Group Type of
JRR; Reviewed:
SPOC 7/21/2009

15 of 70
CM-PE-NN4500
sip. The two new SIP trunk groups are conceptually the same as those defined for trunk
groups 30, 31, 60, and 61 in reference [JRR].
The following shows page 1 for trunk group 32. The Number of Members field defines how
many simultaneous calls are permitted for the trunk group, and can be coordinated with Acme
Packet Net-Net 4500 call admission control features if desired.
add trunk-group 32
Page
1 of
21
TRUNK GROUP
Group Number:
Group Name:
Direction:
Dial Access?
Queue Length:
Service Type:
32
SIP-PSTN-PE
two-way
n
0
public-ntwrk
Group Type: sip

CDR Reports: y
COR: 1
TN: 1
TAC: 132
Outgoing Display? n
Night Service:
Auth Code? n
Signaling Group: 32
Number of Members: 10
The following shows Page 2 for trunk group 32. All parameters shown are default values, except
for the Preferred Minimum Session Refresh Interval, which has been changed from 600 to 900
to avoid unnecessary SIP messaging with the Cisco products used to simulate the SIP Service
Provider. As such, this screen will not be shown for trunk group 62.
add trunk-group 32
Group Type: sip
TRUNK PARAMETERS
Unicode Name: yes
SCCAN? n
Page
2 of
21
Redirect On OPTIM Failure: 5000

Digital Loss Group: 18
Preferred Minimum Session Refresh Interval(sec): 900
The following shows Page 3 for trunk group 32. All parameters shown are at default values. As
such, this screen will not be shown for trunk group 62.
change trunk-group 30
TRUNK FEATURES
ACA Assignment? n
Page
3 of
21
Measured: none
Maintenance Tests? y
Numbering Format: public

UUI Treatment: service-provider
Replace Restricted Numbers? n
Replace Unavailable Numbers? n
JRR; Reviewed:
SPOC 7/21/2009

16 of 70
CM-PE-NN4500
The following shows Page 4 for trunk group 32. All parameters shown are at default values. As
such, this screen will not be shown for trunk group 62. Depending on the Service Provider, it
may be necessary to enter a specific value, such as 101, in the Telephone Event Payload Type
associated with DTMF signaling. Check with the specific service provider. Similarly, some
Service Providers may require that the fields Support Request History and Send Diversion
Header be changed from default values for proper support of redirection features such as
Extension to Cellular or call forwarding off-net.
change trunk-group 30
Page
4 of
21
PROTOCOL VARIATIONS
Mark Users as Phone? n
Prepend '+' to Calling Number? n
Send Transferring Party Information? n
Send Diversion Header? n
Support Request History? y
Telephone Event Payload Type:
The following shows Page 1 for trunk group 62. Remaining pages match trunk group 32.
add trunk-group 62
Page
1 of
21
TRUNK GROUP
Group Number:
Group Name:
Direction:
Dial Access?
Queue Length:
Service Type:
62
Group Type:
SIP-PSTN-ESS-PE
COR:
two-way
Outgoing Display?
n
0
public-ntwrk
Auth Code?
sip
CDR Reports: y
1
TN: 1
TAC: 162
n
Night Service:
n
Signaling Group: 62
Number of Members: 10
3.6. Route Patterns

Route pattern 30 will be used for calls that prefer the SIP trunks at the primary site, but may use
SIP trunks at the secondary site if the SIP trunks at the primary site are busy or failed. Note that
Look-Ahead Routing (LAR) is set to next. As an example of LAR, assume the Acme Packet
4500 at the primary site has just failed, and Communication Manager has not yet marked primary
site trunks out-of-service. Assume that an outbound call is made that chooses this route-pattern.
The call can use LAR for automatic route-advance to complete successfully using the SIP
trunks at the secondary site. Digit manipulation can be performed on the number, if needed. In
the sample configuration, the leading digit (i.e., the 1) is deleted and a 10 digit number is sent.
(This may not be representative of the numbering scheme expected by a production SIP Service
Provider.) In Section 5, the order of the trunks in the route-pattern is varied for specific
verifications.
JRR; Reviewed:
SPOC 7/21/2009

17 of 70
CM-PE-NN4500
change route-pattern 30
Page
Pattern Number: 30 Pattern Name: SIP-PSTN-P
SCCAN? n
Secure SIP? n
Grp FRL NPA Pfx Hop Toll No. Inserted
No
Mrk Lmt List Del Digits
Dgts
1: 32
0
1
2: 31
0
1
3: 30
0
1
4: 60
0
1
5: 61
0
1
6: 62
0
1
1:
2:
3:
4:
5:
6:
BCC VALUE TSC CA-TSC

0 1 2 M 4 W
Request
ITC BCIE Service/Feature PARM
y
y
y
y
y
y
rest
rest
rest
rest
rest
rest
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
n
n
n
n
n
n
n
n
n
n
n
n
1 of
DCS/
QSIG
Intw
n
n
n
n
n
n
IXC
user
user
user
user
user
user
No. Numbering LAR

Dgts Format
Subaddress
next
next
next
next
next
none
Route pattern 60 will be used for calls that prefer the SIP trunks at the secondary site, but may
use the SIP trunks at the primary site, if the SIP trunks at the secondary site are busy or failed.
As with route-pattern 30, LAR is configured to next to allow calls to complete automatically in
failure scenarios. Although not necessary, trunk group 62, associated with the Processor
Ethernet of the ESS, is listed after the C-LAN based trunks at the secondary site, since trunk
group 62 will be out-of-service during normal operation.
change route-pattern 60
Page
Pattern Number: 60 Pattern Name: SIP-PSTN-S
SCCAN? n
Secure SIP? n
Grp FRL NPA Pfx Hop Toll No. Inserted
No
Mrk Lmt List Del Digits
Dgts
1: 61
0
1
2: 60
0
1
3: 62
0
1
4: 31
0
1
5: 30
0
1
6: 32
0
1
1:
2:
3:
4:
5:
6:
BCC VALUE TSC CA-TSC

0 1 2 M 4 W
Request
ITC BCIE Service/Feature PARM
y
y
y
y
y
y
rest
rest
rest
rest
rest
rest
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
JRR; Reviewed:
SPOC 7/21/2009
n
n
n
n
n
n
n
n
n
n
n
n
1 of
DCS/
QSIG
Intw
n
n
n
n
n
n
IXC
user
user
user
user
user
user
No. Numbering LAR

Dgts Format
Subaddress
next
next
next
next
next
none

18 of 70
CM-PE-NN4500
3.7. Administer Public Numbering

The change public-unknown-numbering command may be used to define the format of the
calling party number to be sent. In the bolded rows shown in the abridged output below, all
calls originating from a 5-digit extension beginning with 52 (i.e., 52XXX) will be prefixed with
732852, and a 10 digit calling party number of the form 7328522XXX will be sent, when any of
the SIP trunk groups (30, 31, 32, 60, 61, 62) in the configuration are chosen for the call.
Although not shown, similar configuration covered other telephone extension ranges, such as
51XXX.
change public-unknown-numbering 0
Page
1 of
2
NUMBERING - PUBLIC/UNKNOWN FORMAT
Total
Ext Ext
Trk
CPN
CPN
Len Code
Grp(s)
Prefix
Len
Total Administered: 13
5 5
5
Maximum Entries: 9999
5 52
30
732852
10
5 52
31
732852
10
5 52
32
732852
10
5 52
60
732852
10
5 52
61
732852
10
5 52
62
732852
10
3.8. Configure ARS Analysis For Outbound Call Routing

Location-based routing is configured so that users at different locations that dial the same
telephone number can have calls choose different route-patterns and trunks. In the sample
configuration, users at the primary site that dial PSTN telephone numbers will preferentially use
trunks at the primary site. Similarly, users at the secondary site that dial PSTN telephone
numbers will preferentially use trunks at the secondary site. Upon congestion or failure, calls
can use the alternate sites trunks.
The following screen shows a sample ARS configuration for location 1. If a user at location 1,
such as extension 51003, dials the ARS access code followed by 1-732-852-XXXX, the call will
select route pattern 60.
change ars analysis 1732 location 1
ARS DIGIT ANALYSIS TABLE
Location: 1
Dialed
String
1732852
JRR; Reviewed:
SPOC 7/21/2009
Total
Min Max
11
11
Route
Pattern
60
Call
Type
natl
Node
Num
Page
1 of
Percent Full:
2
1
ANI
Reqd
n

19 of 70
CM-PE-NN4500
The following screen shows a sample ARS configuration for location 3. If a user at location 3,
such as extension 52020, dials the ARS access code followed by 1-732-852-XXXX, the call will
select route pattern 30.
change ars analysis 1732 location 3
ARS DIGIT ANALYSIS TABLE
Location: 1
Dialed
String
1732852
Total
Min Max
11
11
Route
Pattern
30
Call
Type
natl
Node
Num
Page
1 of
Percent Full:
2
1
ANI
Reqd
n
3.9. Configure Incoming Call Handling Treatment For Inbound Digit

Manipulation
The incoming call handling treatment for a trunk group can be used to manipulate the digits
received for an incoming call. In the sample configuration, the number sent from the (simulated)
SIP Service Provider has no direct relationship to the corresponding extension in Communication
Manager. Therefore, all digits are deleted, and the desired Communication Manager
extension is inserted. In the sample configuration, if a PSTN user dials 732-852-1816, the
number 21816 arrives via one of the SIP Trunks (as shown in Figure 2). The incoming call
handling table maps 21816 to 52020, the local extension corresponding to the external PSTN
number. During testing, the number to insert was varied, so that different types of telephones
(e.g., IP, digital) and vector directory numbers (VDN) could be tested.
In reference [JRR], the incoming call handling treatment for trunks 30, 31, 60, and 61 are
presented. The two new SIP trunk groups associated with the Processor Ethernet SIP signaling
groups are similarly configured, as shown below.
change inc-call-handling-trmt trunk-group 32
INCOMING CALL HANDLING TREATMENT
Service/
Number
Number
Del Insert
Feature
Len
Digits
public-ntwrk
5 21816
all 52020
public-ntwrk
5 22940
all 51003
Page
1 of
30
Page
1 of
30
The corresponding configuration for trunk group 62 is shown below.

change inc-call-handling-trmt trunk-group 62
INCOMING CALL HANDLING TREATMENT
Service/
Number
Number
Del Insert
Feature
Len
Digits
public-ntwrk
5 21816
all 52020
public-ntwrk
5 22940
all 51003
JRR; Reviewed:
SPOC 7/21/2009

20 of 70
CM-PE-NN4500
3.10. Announcement-Related Configuration

Section 3.12 of reference [JRR] describes the announcement-related configuration retained in the
verification of these Application Notes. The verifications in Section 5 include calls requesting
these announcements.
3.11. Summarizing VDN and Vector-Related Configuration

Section 3.13 of reference [JRR] describes the VDN and vector-related configuration retained in
the verification of these Application Notes. The verifications in Section 5 include calls to VDNs.
3.12. Summarizing ESS-Related Configuration

Section 3.14 of reference [JRR] describes the ESS-related configuration retained in the
verification of these Application Notes. Product documentation [ESS] should be consulted for
more information on configuring ESS.
3.13. Saving Configuration Changes

The command save translation all can be used to save the configuration. In the sample
configuration, translations were automatically saved and synchronized with each survivable
processor, such as the ESS, on a nightly basis, as a result of the bold parameters in the screen
shown below.
change system-parameters maintenance
MAINTENANCE-RELATED SYSTEM PARAMETERS
Page
1 of
OPERATIONS SUPPORT PARAMETERS

CPE Alarm Activation Level: minor
SCHEDULED MAINTENANCE
Start Time:
Stop Time:
Save Translation:
Update LSP and ESS Servers When Saving Translations:
JRR; Reviewed:
SPOC 7/21/2009
22 : 00
06 : 00
daily
y

21 of 70
CM-PE-NN4500
4. Configure Acme Packet Net-Net Session Directors

This section describes only the changes to the configuration of the Session Directors, compared
to the configuration fully documented in reference [JRR]. Appendix A of reference [JRR]
contains the output of the show running-configuration command from the primary site Session
Director. For additional details on the configuration of the Session Director, refer to [AP1].
The Session Director can be configured via the Acme Packet Command Line Interface (ACLI).
This section assumes the reader is familiar with accessing and configuring the Session Director.
Figure 4 on the following page illustrates a pictorial view of the sample configuration for the
Acme Packet Net-Net 4500 at the primary site. Although specifics such as IP addresses will
vary, the configurations for the Acme Packet Net-Net 4500 at the primary and secondary sites
are conceptually identical, unless noted otherwise. Figure 4 is identical to Figure 3 from
reference [JRR] except that the shaded session-agent with the IP Address of the active S8730
Server has been added to the session-group ENTERPRISE. The configuration of the Session
Director at the secondary site is similar, but of course the appropriate IP Addresses shown in
Figure 1 must be substituted, including the new session-agent with the IP Address of the
Processor Ethernet of the ESS.
JRR; Reviewed:
SPOC 7/21/2009

22 of 70
CM-PE-NN4500
Figure 4: Pictorial View of the Primary Site Session Director Configuration
JRR; Reviewed:
SPOC 7/21/2009

23 of 70
CM-PE-NN4500
4.1. Acme Packet Command Line Interface Summary

The Session Director is configured using the Acme Packet Command Line Interface (ACLI).
The following are the generic ACLI steps for configuring various elements.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Access the console port of the Session Director using a PC and a terminal emulation
program such as HyperTerminal. Use the following settings for the serial port on the PC.
Bits per second: 115200
Data bits: 8
Parity : None
Stop bits: 1
Flow control: None
Log in to the Session Director with the user password.
Enable the Superuser mode by entering the enable command and then the superuser
password. The command prompt will change to include a # instead of a > while in
Superuser mode. This level of system access (i.e., at the acmesystem# prompt) will be
referred to as the main level of the ACLI. Specific sub-levels of the ACLI will then be
accessed to configure specific elements and specific parameters of those elements.
In Superuser mode, enter the configure terminal command. The configure terminal
command is used to access the system level where all operating and system elements may
be configured. This level of system access will be referred to as the configuration level.
Enter the name of an element to be configured (e.g., system).
Enter the name of a sub-element, if any (e.g., phy-interface).
Enter the name of an element parameter followed by its value (e.g., name s0p0).
Enter done to save changes to the element. Use of the done command causes the system
to save and display the settings for the current element.
Enter exit as many times as is necessary to return to the configuration level.
Repeat Steps 4 - 8 to configure all the elements.
Enter exit to return to the main level.
Type save-config to save the entire configuration.
Type activate-config to activate the entire configuration.
After accessing different levels of the ACLI to configure elements and parameters, it is necessary
to return to the main level to run certain tasks such as saving the configuration, activating the
configuration, or rebooting the system.
4.2. System Configuration

The system configuration defines system-wide parameters for the Session Director. The system
configuration documented in Section 4.2 of reference [JRR] applies; no changes are necessary
due to the use of Processor Ethernet SIP trunks in Communication Manager.
4.3. Physical and Network Interfaces

The physical configuration documented in Section 4.3 of reference [JRR] applies. In the sample
configuration, for each Session Director, the Ethernet interface slot 0 / port 0 was connected to
JRR; Reviewed:
SPOC 7/21/2009

24 of 70
CM-PE-NN4500
the external untrusted network, and Ethernet slot 1 / port 0 was connected to the internal
corporate LAN. A network interface was defined for each physical interface to assign it a
routable IP address.
Section 6 documents an issue where the Avaya ESS will expect a response to ICMP ping from
the far-end of SIP trunks, upon activation and control by an ESS running version 5.2 plus SP1.
Although this is a problem to be resolved in a subsequent software load, the commands below
allow a work-around to be implemented on the Acme Packet Session Director at the secondary
site. In the output below, the only changes to the configuration documented in reference [JRR]
are shown in bold. In short, this allows signaling group 62 to remain in-service after an ESS
takes over, if the ESS is running version 5.2 SP1. See Section 5.7.1.2 for verification of this
work-around for signaling group 62.
**sbcsecsite-pri# config t
**sbcsecsite-pri(configure)# system network-interface
**sbcsecsite-pri(network-interface)# select
<name>:<sub-port-id>:
1: wancom1:0 ip=
gw=
2: wancom2:0 ip=
gw=
3: s0p0:0
ip=10.3.3.145
gw=10.3.3.1
4: s1p0:0
ip=2.2.185.145
gw=2.2.185.1
selection: 4
**sbcsecsite-pri(network-interface)# add-hip-ip 2.2.185.145
**sbcsecsite-pri(network-interface)# add-icmp-ip 2.2.185.145
**sbcsecsite-pri(network-interface)# done
network-interface
name
s1p0
sub-port-id
0
description
hostname
ip-address
2.2.185.145
pri-utility-addr
2.2.185.146
sec-utility-addr
2.2.185.147
netmask
255.255.255.0
gateway
2.2.185.1
<text removed>
hip-ip-list
2.2.185.145
ftp-address
icmp-address
2.2.185.145
<text removed>
**sbcsecsite-pri(network-interface)# exit
**sbcsecsite-pri(system)# exit
**sbcsecsite-pri(configure)# exit
**sbcsecsite-pri# save
checking configuration
Save-Config received, processing.
waiting for request to finish
Request to 'SAVE-CONFIG' has Finished,
Save complete
Currently active and saved configurations do not match!
To sync & activate, run 'activate-config' or 'reboot activate'.
*sbcsecsite-pri# activate-config
Activate-Config received, processing.
waiting for request to finish
Request to 'ACTIVATE-CONFIG' has Finished,
Activate Complete
JRR; Reviewed:
SPOC 7/21/2009

25 of 70
CM-PE-NN4500
4.4. Realm
A realm represents a group of related Session Director components. Defining realms allows
flows to pass through a connection point between two networks. Two realms were defined for
the interoperability test. The OUTSIDE realm was defined for the external network and the
INSIDE realm was defined for the internal network. The realm configuration documented in
Section 4.4 of reference [JRR] applies; no changes are necessary due to the use of Processor
Ethernet SIP trunks in Communication Manager.
4.5. SIP Configuration

The SIP configuration (sip-config) defines global system-wide SIP parameters. The SIP
4.6. SIP Interface

The SIP interface (sip-interface) defines the receiving characteristics of the SIP interfaces on the
Session Director. Two SIP interfaces were defined, one for each realm. The SIP interface
4.7. Session Agent

A session agent defines the characteristics of a signaling peer to the Session Director. All the
session-agents previously configured in reference [JRR] continue to apply. In these Application
Notes, a session agent for the Processor Ethernet of the active S8730 Server is added to the
primary site Session Director. Similarly, a session agent for the Processor Ethernet of the S8500
ESS is added to the secondary site Session Director. Key session agent (session-agent)
parameters include:
hostname: Fully qualified domain name or IP address of this SIP peer.
ip-address: The IP address of this SIP peer.
port: The port used by the peer for SIP traffic.
app-protocol: SIP
transport-method: StaticTCP. With static TCP, a TCP connection can be re-used for
multiple sessions. With the alternative DynamicTCP, a new connection must be
established for each session. DynamicTCP also works with Communication Manager,
but since DynamicTCP had already been documented in the compliance testing
performed in reference [AC-HA], static TCP was used in this sample configuration to
show that it is a viable option for interoperability with Communication Manager.
realm-id: The realm id where this peer resides.
description: A descriptive name for the peer.
max-sessions: Although not used in the sample configuration, this parameter can allow
call admission control to be applied for the session agent. For example, in reference
[JSR], the max-sessions parameter is configured to match the number of members in the
corresponding Avaya SIP trunk group. If this is not configured, and the Session Director
sends an INVITE to an Avaya signaling group whose corresponding trunk group has no
available members, Communication Manager will respond with a SIP 503. The Session
JRR; Reviewed:
SPOC 7/21/2009

26 of 70
CM-PE-NN4500
Director will redirect the call to another session agent in the SAG. In the context of a
session agent group whose members are comprised of session agents with different
capabilities, the max-sessions parameter could be used to effect changes to the
distribution of sessions across the various agents.
ping-method: OPTIONS;hops=0 The SIP OPTIONS message will be sent to the peer to
verify that the SIP connection is functional. In addition, this parameter causes the Session
Director to set the Max-Forward field to 0 in outbound OPTIONS pings generated by the
Session Director to this session-agent.
ping-interval: Specifies the interval between SIP OPTIONS pings in seconds. Since
the intent is to monitor the health of the connection, pings may be suppressed if there is
traffic to/from the session-agent that shows the connection is up.
ping-in-service-response-codes: Although not configured as part of the base
verification testing, this parameter can be used to specify the list of response codes that
keep a session agent in-service. By default, any response from the session agent is
enough to keep the session agent in service. If it is desired that only a 200 OK response
is a valid response to OPTIONS to mark the session agent in-service, then 200 can be
entered. Note that Communication Manager will respond to OPTIONS with a 503 in
various conditions where the SIP trunk group corresponding to the SIP signaling group
would not be able to process a call. Examples include when the SIP trunk group is
administratively busied out, or when the trunk group is in-service, but there are no
available resources to handle a call (e.g. all trunk members in use for calls). Other cases
where the Processor Ethernet SIP trunks will respond with a 503 are documented in the
verifications in Section 5.
out-service-response-codes Although not defined in the sample configuration, this
parameter can be used to specify the list of OPTIONS ping response codes that take a
session agent out-of-service.
options trans-timeouts=1 This parameter defines the number of consecutive non-ping
transaction timeouts that will cause the session agent to be marked out-of-service. For
example, with this option set to 1, if an INVITE is sent to an Avaya session agent that is
currently marked in-service, but no response is received resulting in a transaction
timeout, the session agent will be immediately marked out-of-service. In the sample
configuration, where session agent groups are used, this allows future calls to flow to inservice session agents in the group without experiencing a delay due to a transaction
timeout. Note that an explicit error response, such as a 503, is not considered a
transaction timeout.
reuse-connections TCP Enables TCP connection re-use.
tcp-keepalive enabled Enables standard TCP Keep-Alives
tcp-reconn-interval 10 Specifies the idle time, in seconds, before TCP keep-alive
messages are sent.
JRR; Reviewed:
SPOC 7/21/2009

27 of 70
CM-PE-NN4500
The key settings for the session agent to the active S8730 Server are shown below. This
configuration is added on the Acme Packet Session Director at the primary site only.
session-agent
hostname
ip-address
port
state
app-protocol
app-type
transport-method
realm-id
egress-realm-id
description
carriers
allow-next-hop-lp
constraints
max-sessions
< text removed for brevity >
ping-method
ping-interval
ping-send-mode
ping-in-service-response-codes
out-service-response-codes
options
reuse-connections
tcp-keepalive
tcp-reconn-interval
2.2.87.13
2.2.87.13
5060
enabled
SIP
StaticTCP
INSIDE
S8730-PE-Active-IP
enabled
disabled
0
OPTIONS;hops=0
16
keep-alive
trans-timeouts=1
TCP
enabled
10
The key settings for the session agent to the ESS are shown below. This configuration is
added on the Acme Packet Session Director at the secondary site only.
session-agent
hostname
ip-address
port
state
app-protocol
app-type
transport-method
realm-id
egress-realm-id
description
carriers
allow-next-hop-lp
constraints
max-sessions
ping-method
ping-interval
ping-send-mode
ping-in-service-response-codes
out-service-response-codes
options
reuse-connections
tcp-keepalive
tcp-reconn-interval
JRR; Reviewed:
SPOC 7/21/2009
2.2.185.88
2.2.185.88
5060
enabled
SIP
StaticTCP
INSIDE
ESS-PE
enabled
disabled
0
OPTIONS;hops=0
16
keep-alive
trans-timeouts=1
TCP
enabled
10

28 of 70
CM-PE-NN4500
4.8. Session Agent Groups (SAG)

A session agent group is a logical collection of two or more session agents that behave as a
single aggregate entity. In the sample configuration documented in reference [JRR], the two CLANs at each site are configured in a session agent group within the Acme Packet Net-Net 4500
at that site. Although not required, a session agent group is also created for connectivity to the
(simulated) SIP PSTN, for easy adaptation to outside networks with multiple options for next
hops.
For the sample configuration associated with these Application Notes, the session agents
associated with the Avaya Processor Ethernet (defined in Section 4.7) are added. In the
primary site Session Director, the session agent for the active Avaya S8730 Server is added to
the SAG ENTERPRISE. In the secondary site Session Director, the session agent for the
S8500 ESS is added to the SAG ENTERPRISE.
Key session group (session-group) parameters include:
group-name: a unique name for the session agent group.
app-protocol: SIP
strategy: selects the algorithm to use for distribution of traffic among the session agents
in the group. In the sample configuration, a simple roundrobin distribution is selected
for alternating traffic among the session agents. More sophisticated alternatives are
available including leastbusy and propdist (Proportional Distribution) based on
configurable session constraints, and are described in reference [AP1].
dest: The IP Addresses of the session agents that are members of the session agent group.
sag-recursion If enabled, allows re-trying another session agent in the session agent
group after a failure for the previously selected session agent. For example, if an
INVITE message is sent to a session agent, and the session agent does not respond, SAG
recursion allows an INVITE to be automatically directed to another session agent in the
SAG. Those familiar with Communication Manager terminology may benefit from the
following parallel. Conceptually, SAG recursion is similar to Avaya Look-Ahead
Routing, where the SAG is the route-pattern, the session agents are the trunk groups in
the route-pattern and SAG recursion allows LAR to the next trunk in the pattern upon a
failure. Note that this analogy is imperfect in that the Acme Packet Net-Net Session
Director can make decisions about which session-agent in the SAG to choose based on
algorithms that would check usage and load before selecting the next session agent. In
the sample configuration, SAG recursion is enabled for the session agent group
containing the interfaces to Communication Manager.
JRR; Reviewed:
SPOC 7/21/2009

29 of 70
CM-PE-NN4500
Since the session agent group ENTERPRISE already exists, and the only desired change is
for the dest parameter to include the Processor Ethernet entry, the following approach can
be used to change only the dest. The commands shown below were issued on the Session
Director at the primary site. The IP Addresses of the destination session agents can be
included inside parenthesis, separated by spaces, as shown in bold below.
acmesbc-pri# config t
acmesbc-pri(configure)# session-router
acmesbc-pri(session-router)# session-group
acmesbc-pri(session-agent-group)# select
<group-name>:
1: SERV_PROVIDER
2: ENTERPRISE
selection: 2
acmesbc-pri(session-agent-group)# dest (2.2.85.2 2.2.85.20 2.2.87.13)
acmesbc-pri(session-agent-group)# show
session-group
group-name
ENTERPRISE
description
state
enabled
app-protocol
SIP
strategy
RoundRobin
dest
2.2.85.2
2.2.85.20
2.2.87.13
trunk-group
sag-recursion
enabled
stop-sag-recurse
401,407
acmesbc-pri(session-agent-group)#
acmesbc-pri(session-agent-group)# done
JRR; Reviewed:
SPOC 7/21/2009

30 of 70
CM-PE-NN4500
For the Acme Packet Net-Net 4500 at the secondary site, the configuration can also be
updated to add the session agent for the S8500 ESS to the existing ENTERPRISE SAG.
To show an alternative approach, a different syntax is shown in bold below. Rather than
including all the destinations within parenthesis, the dest +2.2.185.88 approach was used to
add the session agent for the ESS to the SAG ENTERPRISE. The before and after
configuration can be observed in the show output below.
sbcsecsite-pri(session-router)# session-group
sbcsecsite-pri(session-agent-group)# select
<group-name>:
1: SERV_PROVIDER
2: ENTERPRISE
selection: 2
sbcsecsite-pri(session-agent-group)# show
session-group
group-name
ENTERPRISE
description
state
enabled
app-protocol
SIP
strategy
RoundRobin
dest
2.2.185.2
2.2.185.20
trunk-group
sag-recursion
enabled
stop-sag-recurse
401,407
last-modified-by
admin@2.2.4.150
last-modified-date
2009-04-14 11:47:50
sbcsecsite-pri(session-agent-group)# dest +2.2.185.88
sbcsecsite-pri(session-agent-group)# show
session-group
group-name
ENTERPRISE
description
state
enabled
app-protocol
SIP
strategy
RoundRobin
dest
2.2.185.2
2.2.185.20
2.2.185.88
trunk-group
sag-recursion
enabled
stop-sag-recurse
401,407
last-modified-by
admin@2.2.4.150
last-modified-date
2009-04-14 11:47:50
sbcsecsite-pri(session-agent-group)#
sbcsecsite-pri(session-agent-group)# done
4.9. SIP Manipulation

SIP manipulations are rules used to modify the SIP messages. For example, SIP manipulations
can be performed to ensure private network topology hiding and confidentiality. The SIP
manipulation configuration documented in Section 4.9 of reference [JRR] applies; no changes
are necessary due to the use of Processor Ethernet SIP trunks in Communication Manager.
4.10. Steering Pools

Steering pools define sets of ports that are used for steering media flows (e.g., RTP) through the
Acme Packet Net-Net 4500. The selected ports are used to modify the SDP to cause receiving
JRR; Reviewed:
SPOC 7/21/2009

31 of 70
CM-PE-NN4500
session agents to direct media to the Acme Packet Net-Net 4500. Two steering pools were
defined, one for each realm. The steering pools documented in Section 4.10 of reference [JRR]
apply; no changes are necessary due to the use of Processor Ethernet SIP trunks in
Communication Manager. Consult reference [AP1] for more information, including a means to
use steering pool configuration for call admission control.
4.11. Local Policy

Local policy controls the routing of SIP calls from one realm to another. The local policies
documented in Section 4.11 of reference [JRR] apply; no changes are necessary due to the use of
Processor Ethernet SIP trunks in Communication Manager.
JRR; Reviewed:
SPOC 7/21/2009

32 of 70
CM-PE-NN4500
5. Verifications
This section illustrates expected behaviors and sample results, focusing on calls using the
Processor Ethernet SIP trunks. Reference [JRR] can be consulted for a robust set of related call
verifications using the C-LAN SIP trunks, including failure scenarios. Section 6 documents test
observations that resulted in product modification requests.
5.1. Normal Operation

When all components are functioning normally, calls are processed by the active S8730 Server at
the primary site. Inbound and outbound calls can use SIP Trunks at either site, subject to routing
rules and efficient allocation of resources. In normal operation, the Processor Ethernet of the
Avaya S8730 Servers can process calls, but the Processor Ethernet of the ESS will not participate
in SIP call traffic.
The following screen, taken from the active S8730 Server during normal operation, shows that
cluster 1, the S8730 Server pair, controls both the primary and secondary site. The Connected
Clus(ter) IDs shows that cluster 2, the S8500 ESS, can be reached by the IPSIs.
status ess port-networks
Cluster ID 1
ESS PORT NETWORK INFORMATION
Com
PN Num
1
1
2
2
Intf
Loc
1A01
2A01
Intf
Type
IPSI
IPSI
Port IPSI
Ntwk Gtway
Ste Loc
up
1A01
up
2A01
Pri/
Sec
Loc
1A01
2A01
Pri/
Cntl Connected
Sec
Clus Clus(ter)
State
ID IDs
actv-aa
1
1
2
actv-aa
1
1
2
The following screen, taken from the active S8730 Server during normal operation, shows that
the primary and secondary site IPSIs are controlled by the S8730 Server and in-service".
list ipserver-interface
IP SERVER INTERFACE INFORMATION
Port
Ntwk
Num
---1
Pri/
Sec
Bd Loc
-----1A01
2A01
Primary/
Secondary
IP Address
--------------2.2.185.9
Primary/
Secondary
Host Name
---------------2.2.185.9
Primary/
Secondary Serv
DHCP ID
State
--------- ----ipsi-A01a IN
Control
State
------actv-aa
2.2.85.9
2.2.85.9
ipsi-A02a
actv-aa 0.0.0.0
JRR; Reviewed:
SPOC 7/21/2009
IN

State Of
Health
C P E G
------0.0.0.0
33 of 70
CM-PE-NN4500
The following screen, taken during normal operation, shows that the ESS with IP Address
2.2.185.88 is registered with up-to-date translations.
list survivable-processor
Name
SURVIVABLE PROCESSORS
IP Address
Reg Act
Type
ESSCid002Sid003 ESS S
S83LSP-in-G250 LSP
S83LSP-in-G700 LSP
2.2.185.88
2.2.25.88
2.2.1.88
y
y
y
n
n
n
Translations
Updated
22:00 6/18/2009
22:00 6/18/2009
22:00 6/18/2009
Net
Rgn
1
2
4
In normal operation, when all network elements are functioning, and the active Avaya S8730
is processing all calls, incoming and outgoing SIP trunk calls to the enterprise can use SIP trunk
groups 30, 31, 32, 60, or 61. Table 2 shows the expected states during normal operation. Trunk
group 62 will not be used during normal operation because Communication Manager trunk
group 62 will be out-of-service (OOS). Also, the secondary site Acme Packet session agent to
the ESS will be out-of-service.
Avaya Trunk
Group (TG) &
Service State via
Avaya S8730
Servers
30 (C-LAN)
in-service
31 (C-LAN)
in-service
32 (S8730 PE)
in-service
60 (C-LAN)
in-service
61 (C-LAN)
in-service
62 (S8730 PE)
Out-of-service /
Far-End
Reason for
Avaya
Service State
OK response
to OPTIONS
OK response
to OPTIONS
OK response
to OPTIONS
OK response
to OPTIONS
OK response
to OPTIONS
S8730 PE is
not a session
agent in the
secondary site
NN4500
Primary Site
NN4500
Session Agent
& Service
State
SA 2.2.85.2
In-service
SA 2.2.85.20
In-service
SA 2.2.87.13
In-service
N/A
N/A
N/A
Secondary
Site NN4500
Session Agent
& Service
State
N/A
N/A
N/A
SA 2.2.185.2
In-service
SA 2.2.185.20
In-service
SA 2.2.185.88
Out-of-Service
Reason for
Acme Packet
Service State
Where
Applicable
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
No SIP response from
ESS at secondary site,
ESS not active
Table 2: Normal Operation, All Network Elements Functioning
JRR; Reviewed:
SPOC 7/21/2009

34 of 70
CM-PE-NN4500
During normal operation, signaling group 62 will be marked for bypass, as shown below. The
Session Director at the secondary site does not respond at the SIP level to the OPTIONS message
sent by the active S8730 Server procr, the near-end of signaling group 62.
status signaling-group 62
STATUS SIGNALING GROUP
Group ID:
Group Type:
Signaling Type:
Group State:
62
sip
facility associated signaling
far-end bypass
Active NCA-TSC Count: 0

Active CA-TSC Count: 0
During normal operation, the members of trunk group 62 will all be out-of-service/far-end as
shown in the abridged output below. Signaling group 62 and trunk group 62 will only be inservice when the ESS is active (e.g., controlling a port network or gateway).
status trunk 62
Page
TRUNK GROUP STATUS

Member
Port
Service State
0062/001
0062/002
0062/003
0062/004
T00140
T00141
T00142
T00143
OOS/FE-idle
OOS/FE-idle
OOS/FE-idle
OOS/FE-idle
Mtce Connected Ports

Busy
no
no
no
no
5.1.1. Incoming Calls from PSTN Arriving via SIP Trunk to Primary Site
Using the sample configuration, incoming PSTN calls arriving via the primary site will be
distributed in a round-robin fashion to the three session agents in the SAG ENTERPRISE.
These include the two C-LAN session agents previously illustrated in reference [JRR] as well as
the new session agent corresponding to the Processor Ethernet of the active S8730 Server.
The following show sipd agent command was run on the primary Acme Packet Net-Net 4500
after three inbound PSTN calls via the primary site Session Director. All session agents are inservice. Session agent 10.3.3.40 (the SIP Service Provider) shows 3 Active Inbound sessions.
The three session agents that are members of the SAG ENTERPRISE each show one Active
Outbound session, as a result of the round-robin distribution.
acmesbc-pri# show sipd agent
10:27:29-34 (recent)
----- Inbound ----Session Agent
Active Rate ConEx
10.3.3.40
I
3
0.0
0
2.2.85.2
I
0
0.0
0
2.2.85.20
I
0
0.0
0
2.2.87.13
I
0
0.0
0
---- Outbound ----- -- Latency -Max

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.010 0.012
1
1
0.0
0 0.118 0.118
1
1
0.0
0 0.118 0.118
1
1
0.0
0 0.002 0.003
1
The following trace output shows a call incoming on signaling group 32 / trunk group 32 from
PSTN telephone 732-852-2550. The incoming call handling table for trunk group 32 mapped the
received number (21816) to extension 52020. Extension 52020 is an IP Telephone with IP
Address 2.2.1.109 in Region 3. Initially, the IP Media Processor in region 3 (2.2.26.4) is used,
JRR; Reviewed:
SPOC 7/21/2009

35 of 70
CM-PE-NN4500
but as can be seen in the final trace output, once the call is answered, the final RTP media path is
ip-direct from the IP Telephone (2.2.1.109) to the inside of the Acme Packet Net-Net 4500
at the primary site (2.2.85.45).
list trace tac 132
Page
LIST TRACE
time
10:59:36
10:59:36
10:59:36
10:59:36
10:59:36
10:59:36
10:59:36
10:59:42
10:59:42
10:59:42
data
Calling party trunk-group 32 member 1 cid 0xa97
Calling Number & Name 7328522496 NO-CPName
active trunk-group 32 member 1 cid 0xa97
dial 52020
ring station
52020 cid 0xa97
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.26.4]:13828
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49160
rgn:3 [2.2.26.4]:13820
active station
52020 cid 0xa97
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49160
rgn:3 [2.2.1.109]:19036
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.85.45]:49160
The following portion of a Wireshark trace shows an incoming call using the S8730 Processor
Ethernet. In frame 49, the Service Provider delivers the inbound INVITE to the Session Director
at the primary site. In Frame 51, the Session Director sends the INVITE to the S8730 Processor
Ethernet. The call proceeds as usual using the S8730 Processor Ethernet.
JRR; Reviewed:
SPOC 7/21/2009

36 of 70
CM-PE-NN4500
In the next example, a call arrives via the Processor Ethernet SIP trunk (32), and the call is
directed to call vector 1 via a VDN (x51081) that plays an announcement (x22232), and then
collects and verifies password digits from a caller. This type of call verifies proper collection of
DTMF via RFC 2833, and also illustrates the Communication Manager audio group concept that
allows announcements to be sourced from the local Avaya gateway. Audio groups enable
efficient utilization of resources, and also provide redundancy benefits.
From the bolded rows, note that the tone receiver as well as the announcements are sourced from
a board in the 2A carrier at the primary site.
list trace tac 132
Page
LIST TRACE
time
11:05:34
11:05:34
11:05:34
11:05:34
11:05:34
11:05:34
11:05:36
11:05:36
11:05:36
11:05:50
11:05:50
data
Calling party trunk-group 32 member 1 cid 0xa9f
active trunk-group 32 member 1 cid 0xa9f
dial 51081
ring vector 1
cid 0xa9f
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49172
rgn:3 [2.2.26.4]:13956
tone-receiver
02AXX04 cid 0xa9f
active announcement
22232 cid 0xa9f
hear audio-group 1 board 02A08 ext 22232 cid 0xa9f
active announcement
22234 cid 0xa9f
hear audio-group 1 board 02A08 ext 22234 cid 0xa9f
JRR; Reviewed:
SPOC 7/21/2009

37 of 70
CM-PE-NN4500
In the next example, a call arrives via the Processor Ethernet site SIP trunk (32), and the call is
directed to Avaya call vector 3 via a VDN (x51082) that plays an announcement (x22555) and
collects digits for call routing. This also shows the locally-sourced announcements, verifies
DTMF collection using RFC 2833, and illustrates a call that arrives via the primary site
Processor Ethernet SIP trunk, but connects with a user (x51003) at the secondary site.
list trace tac 132
Page
LIST TRACE
time
11:09:13
11:09:13
11:09:13
11:09:13
11:09:13
11:09:13
11:09:15
11:09:15
11:09:15
11:09:26
11:09:26
11:09:26
11:09:31
11:09:31
11:09:31
data
Calling party trunk-group 32 member 1 cid 0xaa3
active trunk-group 32 member 1 cid 0xaa3
dial 51082
ring vector 3
cid 0xaa3
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49178
rgn:3 [2.2.26.4]:14040
tone-receiver
02B0108 cid 0xaa3
active announcement
22555 cid 0xaa3
hear audio-group 1 board 02A08 ext 22555 cid 0xaa3
dial 51003
ring station
51003 cid 0xaa3
G711MU ss:off ps:20
rgn:1 [2.2.185.200]:3472
rgn:3 [2.2.26.4]:14048
active station
51003 cid 0xaa3
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49178
rgn:1 [2.2.185.200]:3472
G711MU ss:off ps:20
rgn:1 [2.2.185.200]:3472
rgn:3 [2.2.85.45]:49178
JRR; Reviewed:
SPOC 7/21/2009

38 of 70
CM-PE-NN4500
5.1.2. Outgoing Call to PSTN from Primary Site

The following trace shows an outbound ARS call from IP Telephone x52020 to the PSTN
number 17328522550. The call is routed based on the location of the originator to route pattern
30, which contains trunk group 32. The call initially uses a media processor in region 3
(2.2.26.4), but after the call is answered, the call is shuffled to become an ip-direct
connection between the IP Telephone (2.2.1.109) and the Acme Packet Session Director at the
primary site (2.2.85.45).
list trace tac 132
Page
LIST TRACE
time
16:25:53
16:25:53
16:25:53
16:25:53
16:25:53
16:25:53
16:25:53
16:25:59
16:25:59
16:25:59
data
dial 9917328522550 route:ARS
route-pattern 30 preference 1 cid 0xa0b
seize trunk-group 32 member 2 cid 0xa0b
Setup digits 7328522550
Calling Number & Name 7328522020 John Public
Proceed trunk-group 32 member 2 cid 0xa0b
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49154
rgn:3 [2.2.26.4]:2764
active trunk-group 32 member 2 cid 0xa0b
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:58096
rgn:3 [2.2.85.45]:49154
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49154
rgn:3 [2.2.1.109]:58096
The following portion of a Wireshark trace shows an outgoing call using the S8730 Processor
Ethernet. In frame 1670, Communication Manager uses the S8730 Processor Ethernet to send an
INVITE to the primary site Session Director. In frame 1673, the Session Director sends the
INVITE on to the Service Provider. The call proceeds as usual using the S8730 Processor
Ethernet.
JRR; Reviewed:
SPOC 7/21/2009

39 of 70
CM-PE-NN4500
Outbound calls were also made to PSTN destinations requiring a log-in with password, such as a
messaging system, to verify that DTMF was working properly in the outbound direction. The
following is an example Communication Manager trace using the Processor Ethernet SIP trunk
62. The call is placed from an H.323 IP Telephone (x52020). When the call is answered, the
call is shuffled to ip-direct, from the IP Telephone (2.2.1.109) to the primary site Session
Director (2.2.85.45). After the user hears the prompt for password input, and begins to enter
digits at 10:07:18, the media shuffles back to the Media Processor serving region 3 at 2.2.26.4.
Once user digit entry ends, and the user in this case is listening to a voice message, the media
shuffles back to ip-direct, as can be seen at 10:07:56. This behavior is not unique to calls using
Processor Ethernet SIP Trunks.
list trace tac 132
Page
LIST TRACE
time
10:07:13
10:07:13
10:07:13
10:07:13
10:07:13
10:07:13
10:07:13
10:07:13
10:07:13
10:07:13
10:07:18
10:07:18
10:07:56
10:07:56
data
route-pattern 30 preference 1 cid 0xa81
seize trunk-group 32 member 13 cid 0xa81
Proceed trunk-group 32 member 13 cid 0xa81
active trunk-group 32 member 13 cid 0xa81
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49154
rgn:3 [2.2.26.4]:13388
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.85.45]:49154
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49154
rgn:3 [2.2.1.109]:19036
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.26.4]:13416
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49154
rgn:3 [2.2.26.4]:13424
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.85.45]:49154
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49154
rgn:3 [2.2.1.109]:19036
5.1.3. Incoming Calls from PSTN Arriving Via PSTN to Secondary Site
Using the sample configuration, incoming PSTN calls arriving via the secondary site will be
distributed in a round-robin fashion to the session agents in the SAG ENTERPRISE defined in
the secondary site Session Director. These include the two C-LAN session agents previously
verified in reference [JRR] as well as the new session agent corresponding to the Processor
Ethernet of the ESS. Under normal operation, as noted previously, the session agent
corresponding to the Processor Ethernet of the ESS will be out-of-service from the point of view
of the secondary site Session Director. Therefore, incoming calls from the PSTN arriving via the
secondary site are identical to those shown in Section 5.1.3 of reference [JRR].
JRR; Reviewed:
SPOC 7/21/2009

40 of 70
CM-PE-NN4500
To illustrate the behavior, the following show sipd agent command was run on the secondary
Acme Packet Net-Net 4500 after three inbound PSTN calls via the secondary site. Observe that
session agent 2.2.185.88 is out of service. Session agent 10.3.3.1 (the SIP Service Provider)
shows 3 Active Inbound sessions. Session agent 2.2.185.2 show two Active Outbound sessions,
and session agent 2.2.185.20 shows one Active Outbound session, as a result of the round-robin
distribution, and the bypass of the out-of-service session agent in the SAG ENTERPRISE.
sbcsecsite-pri# show sipd agent
11:33:28-56 (recent)
Active Rate ConEx
10.3.3.1
I
3
0.0
0
2.2.185.2
I
0
0.0
0
2.2.185.20
I
0
0.0
0
2.2.185.88
O
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.004 0.005
1
2
0.0
0 0.098 0.118
1
1
0.0
0 0.118 0.118
1
0
0.0
0 0.000 0.000
0
5.1.4. Outgoing Call to PSTN from Secondary Site

During normal operation, outgoing calls to the PSTN via the secondary site are identical to those
shown in Section 5.1.4 of reference [JRR]. Since the Processor Ethernet of the S8500 ESS is not
chosen for outbound SIP trunk calls during normal operation, there is no change to the results
compared to reference [JRR].
5.2. Enterprise IP WAN Failure Isolating Secondary Site (ESS Controls

Secondary Site Only)
This section illustrates behavior when the enterprise IP network that allows the secondary site to
communicate with other sites is down. Refer to Induced Failure Reference Number 1 in
Figure 1. Since the secondary site is isolated from the primary site, the ESS in the secondary
site controls the Avaya G650 Media Gateway in the secondary site. The primary site has not
experienced a failure, and therefore the primary site active S8730 Server still controls the
primary site Avaya G650 Media Gateway.
The following screen, taken from the ESS, shows that the ESS is controlling port network 1 only.
Port
Ntwk
Num
---1
2
Pri/
Sec
Bd Loc
-----1A01
2A01
Primary/
Secondary
IP Address
--------------2.2.185.9
2.2.85.9
JRR; Reviewed:
SPOC 7/21/2009
Primary/
Secondary
Host Name
---------------2.2.185.9
2.2.85.9
Primary/
Secondary Serv
DHCP ID
State
--------- ----ipsi-A01a IN
ipsi-A02a OUT
Control
State
------actv-aa
active

State Of
Health
C P E G
------0.0.0.0
0.1.1.0
41 of 70
CM-PE-NN4500
Table 3 shows the expected states from the point of view of the active Avaya S8730 Server,
when the WAN is down, and the primary site is controlled by the active Avaya S8730 Server,
and the secondary site is controlled by the Avaya ESS. Incoming and outgoing SIP trunk calls at
the primary site can use SIP trunk groups 30, 31, and 32.
Avaya Trunk
Group (TG) &
Service State
via Avaya
S8730 Server
30 (C-LAN)
in-service
31 (C-LAN)
in-service
32 (S8730 PE)
in-service
60 (C-LAN)
Out-of-service /
Near-End
61 (C-LAN)
Out-of-service /
Near-End
62 (S8730 PE)
Out-of-service /
Far-End
Reason for Primary Site

Avaya
NN4500
Service State
Session
via S8730
Agent &
Server
Service State
OK response
SA 2.2.85.2
to OPTIONS
In-service
OK response SA 2.2.85.20
to OPTIONS
In-service
OK response SA 2.2.87.13
to OPTIONS
In-service
S8730 not
N/A
controlling
C-LANs
cabinet
S8730 not
N/A
controlling
C-LANs
cabinet
No response
N/A
from far-end,
the secondary
site NN4500
Secondary Site
NN4500
Session Agent
& Service
State
N/A
SA 2.2.185.2
In-service
Reason for
Acme Packet
Service State
Where
Applicable
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS
SA 2.2.185.20
In-service
Any SIP response to

OPTIONS
SA 2.2.185.88
In-Service
Any SIP response to

OPTIONS (from
active ESS Processor
Ethernet)
N/A
N/A
Table 3: S8730 Controls Primary Site, ESS Controls Secondary Site, S8730 States
JRR; Reviewed:
SPOC 7/21/2009

42 of 70
CM-PE-NN4500
Table 4 shows the expected states from the point of view of the Avaya ESS, when the WAN is
down, and the primary site is controlled by the active Avaya S8730 Server, and the secondary
site is controlled by the ESS. Incoming and outgoing SIP trunk calls at the secondary site can use
SIP trunk groups 60, 61, and 62. Trunk group 62 and the corresponding session agent in the
secondary site Session Director come in-service when the Avaya ESS becomes active.
Avaya Trunk
Group (TG) &
Service State
via Avaya
S8500 ESS
30 (C-LAN)
Out-of-service /
Near-End
31 (C-LAN)
Out-of-service /
Near-End
32 (S8500
procr)
Out-of-service /
Far-End
60 (C-LAN)
In-service
61 (C-LAN)
In-service
62 (S8500
procr)
In-service
Reason for
Avaya
Service State
via ESS
ESS not
controlling
cabinet
containing
C-LAN
ESS not
controlling
cabinet
containing
C-LAN
No response
from primary
site Session
Director
OK response
to OPTIONS
OK response
to OPTIONS
See Section
5.7
Primary Site
NN4500
Session Agent
& Service
State
SA 2.2.85.2
In-service
Secondary
Site NN4500
Session Agent
& Service
State
N/A
Reason for
Acme Packet
Service State
Where
Applicable
Any SIP response to
OPTIONS
SA 2.2.85.20
In-service
N/A
Any SIP response to

OPTIONS
SA 2.2.87.13
In-service
N/A
Any SIP response to

OPTIONS
N/A
SA 2.2.185.2
In-service
SA 2.2.185.20
In-service
SA 2.2.185.88
In-Service
Any SIP response to

OPTIONS
Any SIP response to
OPTIONS
Any SIP response to
OPTIONS (from
active ESS Processor
Ethernet)
N/A
N/A
Table 4: S8730 Controls Primary Site, ESS Controls Secondary Site, ESS States
5.2.1. Incoming Call from PSTN to Primary Site

Since the primary site remains under the control of the Avaya S8730 Server, incoming calls from
the SIP Trunks at the primary site to primary site users are identical to those illustrated in Section
5.1.1. These call traces will not be repeated here. See Section 5.2.1 of reference [JRR] for Dial
Plan Transparency traces and considerations.
JRR; Reviewed:
SPOC 7/21/2009

43 of 70
CM-PE-NN4500
The following screen, taken from the primary site Session Director in this condition, shows that
all the session agents configured on the primary site Session Director are in-service.
10:12:11-46 (recent)
Active Rate ConEx
10.3.3.40
I
0
0.0
0
2.2.85.2
I
0
0.0
0
2.2.85.20
I
0
0.0
0
2.2.87.13
I
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.006 0.007
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.002 0.002
1
5.2.2. Outgoing Call to PSTN from Primary Site

Since the primary site remains under the control of the Avaya S8730 Server, calls from primary
site users are identical to those illustrated in Section 5.1.2. Traces will not be repeated here.
5.2.3. Incoming Call from PSTN to Secondary Site

When the WAN initially fails (i.e., during the configurable IPSI no-service timeout period),
the Avaya G650 Media Gateway at the secondary site will not be controlled by any server.
Therefore, the secondary site Session Director will mark the session agents to the C-LANs at the
secondary site out-of-service. Until the ESS is activated, taking over the Avaya G650 Media
Gateway at the secondary site, all the session agents in the ENTERPRISE SAG at the
secondary site will be out-of-service, as shown in the screen below. In the state shown below, if
a call is delivered by the PSTN to the secondary site Session Director, the Session Director will
return a 503 Service Unavailable, and the fail-over capability of the service provider network can
redirect the call to the primary site if applicable.
11:08:28-56 (recent)
Active Rate ConEx
10.3.3.1
I
0
0.0
0
2.2.185.2
O
0
0.0
0
2.2.185.20
O
0
0.0
0
2.2.185.88
O
0
0.0
0
JRR; Reviewed:
SPOC 7/21/2009

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.004 0.004
1
0
0.0
1 0.000 0.000
1
0
0.0
0 0.000 0.000
1
0
0.0
0 0.000 0.000
0

44 of 70
CM-PE-NN4500
Once the ESS is active, an incoming call from the PSTN via the secondary site can be distributed
by the secondary site Session Director to the Processor Ethernet interface of the ESS (as well as
the C-LAN session agents in the secondary site). That is, the Session Director will mark the
session agent to the ESS in-service, since the ESS begins responding to the SIP OPTIONS
messages from the Session Director. The following screen illustrates the agent state. The screen
below and the previous screen were taken during the testing of ESS take-over of the secondary
site, and the timestamps are therefore relevant. At roughly 11:08, a failure to the WAN was
induced. By roughly 11:12 (IPSI no-service timer was configured for 3 minutes), all the session
agents at the secondary site were in-service, under the control of the ESS.
11:12:42-40 (recent)
Active Rate ConEx
10.3.3.1
I
0
0.0
0
2.2.185.2
I
0
0.0
0
2.2.185.20
I
0
0.0
0
2.2.185.88
I
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.004 0.004
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.083 0.118
1
0
0.0
0 0.002 0.002
0
At this point, in Communication Manager running on the ESS, signaling groups 60 through 62,
and trunk groups 60 through 62, are in-service and eligible for call activity.
The following screen shows an example distribution of 3 inbound calls via the secondary site
Session Director, when the ESS is controlling cabinet 1. Note that 3 inbound calls are active
from the PSTN (10.3.3.1), and these three calls have been distributed in a round-robin fashion to
each of the three session agents, including the ESS Processor Ethernet (2.2.185.88).
sbcsecsite-pri# show sipd ag
11:36:10-38 (recent)
Active Rate ConEx
10.3.3.1
I
3
0.0
0
2.2.185.2
I
0
0.0
0
2.2.185.20
I
0
0.0
0
2.2.185.88
I
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.004 0.005
1
1
0.0
0 0.118 0.118
1
1
0.0
0 0.119 0.120
1
1
0.0
0 0.002 0.002
1
The following portion of a Wireshark trace shows an incoming call using the S8500 ESS
Processor Ethernet. In frame 16472, the Service Provider delivers the inbound INVITE to the
Session Director at the secondary site. In Frame 16474, the Session Director sends the INVITE
to the S8500 ESS Processor Ethernet. The call proceeds as usual using the ESS Processor
Ethernet.
JRR; Reviewed:
SPOC 7/21/2009

45 of 70
CM-PE-NN4500
The following status screen, taken from the ESS, shows an active call via the Processor Ethernet
of the ESS, using signaling group and trunk group 62.
status trunk 62
Member
Port
0062/001
0062/002
0062/003
0062/004
T00140
T00141
T00142
T00143
Page
TRUNK GROUP STATUS

Service State
Mtce Connected Ports
Busy
in-service/active no
S00000
in-service/idle
no
in-service/idle
no
in-service/idle
no
5.2.4. Outgoing Call to PSTN from Secondary Site

Once the ESS is active, signaling group 62 using the Processor Ethernet of the ESS will be inservice. Therefore, trunk group 62 may be chosen for outbound calls, subject to normal routing
decisions.
The following trace, taken from the ESS, shows an outbound ARS call from IP Telephone
x51003 to the PSTN number 17328522550. (Note that this is the same telephone number called
in Section 5.1.2). The call is routed based on the location of the originator to route pattern 60,
which contains trunk group 62. (For this call, trunk group 62 was listed first in route pattern 60
for illustration purposes). The call initially uses a media processor in region 1 (2.2.185.4), but
after the call is answered, the call is shuffled to become an ip-direct connection between the
IP Telephone (2.2.185.200) and the Acme Packet at the secondary site (2.2.185.145).
list trace station 51003
l
LIST TRACE
time
data
09:49:44
active station
51003 cid 0x26
09:49:44
G711MU ss:off ps:20
rgn:1 [2.2.185.200]:3472
rgn:1 [2.2.185.4]:2464
09:49:47
09:49:47
term trunk-group 62
cid 0x26
09:49:48
09:49:48
route-pattern 60 preference 1 cid 0x26
09:49:48
seize trunk-group 62 member 2 cid 0x26
09:49:48
09:49:48
Calling Number & Name 7328521003 Peter Parker
09:49:49
Proceed trunk-group 62 member 2 cid 0x26
09:49:49
G711MU ss:off ps:20
rgn:1 [2.2.185.145]:49170
rgn:1 [2.2.185.4]:2472
09:49:54
active trunk-group 62 member 2 cid 0x26
09:49:54
G711MU ss:off ps:20
rgn:1 [2.2.185.200]:3472
rgn:1 [2.2.185.145]:49170
09:49:54
G711MU ss:off ps:20
rgn:1 [2.2.185.145]:49170
rgn:1 [2.2.185.200]:3472
Page
The following portion of a Wireshark trace shows an outgoing call using the S8500 ESS
Processor Ethernet. In frame 75, Communication Manager running on the ESS sends an INVITE
via the S8500 Processor Ethernet to the secondary site Session Director. In frame 78, the
JRR; Reviewed:
SPOC 7/21/2009

46 of 70
CM-PE-NN4500
Session Director sends the INVITE to the Service Provider. The call proceeds as usual using the
S8500 ESS Processor Ethernet.
If the configuration were different, such that a call from the secondary site user was directed to a
routing pattern where the SIP trunks at the primary site were listed before the SIP trunks at the
secondary site, see Section 6.
5.3. Avaya S8730 Servers Off-Line (ESS Controls Primary and

Secondary Site)
This section pertains to cases where both Avaya S8730 Servers are off-line. Refer to induced
failure reference number 4 in Figure 1. For the S8730 Server interchange scenario (i.e., where
the standby server becomes the active and vice-versa), see Section 5.8.
From the time the failure is induced until the expiration of the IPSI no-service timer governing
ESS fail-over, all inbound calls to the enterprise will fail (unless alternate routing is provided via
the service provider when both locations in Figure 1 are non-responsive). The Session Directors
at both sites will get no response from Communication Manager. As described in Section 1.3, an
inbound call can result in a SIP 408 back to the SIP Service Provider, due to a transaction
timeout, or a 503 back to the SIP Service Provider, once the session agents are marked out-ofservice.
After the IPSI no-service timer expires, the ESS will take control over the Avaya G650 Media
Gateways at both the primary and secondary sites. The following screen, taken from the active
ESS, shows that the ESS is control of both cabinets.
Port
Ntwk
Num
---1
Pri/
Sec
Bd Loc
-----1A01
2A01
Primary/
Secondary
IP Address
--------------2.2.185.9
Primary/
Secondary
Host Name
---------------2.2.185.9
Primary/
Secondary Serv
DHCP ID
State
--------- ----ipsi-A01a IN
Control
State
------actv-aa
2.2.85.9
2.2.85.9
ipsi-A02a
actv-aa 0.0.0.0
JRR; Reviewed:
SPOC 7/21/2009
IN

State Of
Health
C P E G
------0.0.0.0
47 of 70
CM-PE-NN4500
The following screen, taken from the primary site Session Director, shows that the two C-LAN
session agents at the primary site are back in-service, since the ESS is in control of the cabinets
at both sites. In this case, the session agent to the S8730 Processor Ethernet is out-of-service,
since the type of failure induced was a failure of the S8730 server pair.
08:33:40-58 (recent)
Active Rate ConEx
10.3.3.40
I
0
0.0
0
2.2.85.2
I
0
0.0
0
2.2.85.20
I
0
0.0
0
2.2.87.13
O
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.011 0.027
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.000 0.000
1
The following screen, taken from the secondary site Session Director, shows that the two C-LAN
session agents at the secondary site are also in-service, since the ESS is in control of both sites.
The session agent to the S8500 ESS Processor Ethernet is also in-service, since the ESS responds
to SIP OPTIONS from the secondary site Session Director when it is active. In the verification
testing, the session agent for the ESS would typically come in-service slightly before the session
agents for the C-LANs newly controlled by the ESS. As can be seen in the screen below, an
inbound PSTN call via the secondary site Session Director is active, having been delivered to
Communication Manager via the ESS Processor Ethernet SIP Trunk (62).
09:32:54-36 (recent)
Active Rate ConEx
10.3.3.1
I
1
0.0
0
2.2.185.2
I
0
0.0
0
2.2.185.20
I
0
0.0
0
2.2.185.88
I
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.004 0.005
2
0
0.0
0 0.072 0.098
0
0
0.0
0 0.072 0.098
0
1
0.0
0 0.002 0.002
1
The Communication Manager call traces for incoming calls to the secondary site when the ESS
is in control of both sites are the same as when the ESS is control of only the secondary site
(Section 5.2), so these call traces are omitted. Of course, since the system is whole in this
case, with all sites controlled by the ESS, there is no need for Dial Plan Transparency to
complete a call that arrives via one site, but is directed to a user at the other. The following
screen shows three active inbound calls from the PSTN that have been evenly distributed by the
secondary site Session Director to the two C-LANs and the Processor Ethernet of the ESS.
09:45:30-42 (recent)
Active Rate ConEx
10.3.3.1
I
3
0.0
0
2.2.185.2
I
0
0.0
0
2.2.185.20
I
0
0.0
0
2.2.185.88
I
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.004 0.005
2
1
0.0
0 0.118 0.118
1
1
0.0
0 0.118 0.119
1
1
0.0
0 0.002 0.003
1
The following screen shows three active inbound calls from the PSTN to the primary site Session
Director. Since the session agent for the active S8730 Processor Ethernet is out-of-service, the
calls are distributed via the two C-LANs at the primary site. The Communication Manager call
JRR; Reviewed:
SPOC 7/21/2009

48 of 70
CM-PE-NN4500
traces for incoming calls to the primary site when the ESS is in control are the same as when the
primary site S8730 Server is in control (Section 5.1), so these call traces are omitted.
09:00:24-42 (recent)
Active Rate ConEx
10.3.3.40
I
3
0.0
0
2.2.85.2
I
0
0.0
0
2.2.85.20
I
0
0.0
0
2.2.87.13
O
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.007 0.008
1
1
0.0
0 0.097 0.118
1
2
0.0
0 0.107 0.118
1
0
0.0
0 0.000 0.000
1
Outbound calls to the PSTN when the ESS is in control of both sites are eligible to use trunk
groups 30, 31, 60, 61, and 62 (see Section 5.7 and Section 6 if using Service Pack 1). Outbound
call traces will look the same as those previously illustrated and are therefore omitted. In the
steady state, trunk group 32 on the active ESS will be marked out-of-service by Communication
Manager maintenance (i.e., no response from primary site Session Director, which does not have
a session agent for the ESS procr). Therefore, trunk group 32 will not be chosen for outbound
calls. Until trunk group 32 is marked out-of-service by Communication Manager maintenance,
any outbound calls directed to a route-pattern containing trunk group 32 can complete to
alternate trunks in the route-pattern via Look-Ahead Routing.
5.3.1. Recovery of the S8730 Servers while the ESS Retains Control
Now assume that the problem that caused the S8730 Servers to be off-line is resolved. Although
the active S8730 Server is not controlling any cabinet, it can still respond to SIP OPTIONS from
the primary site Session Director. Since the S8730 Server is not in control of any resources, the
response to SIP OPTIONS is a 503 Service Unavailable. Recall that by default, the Session
Director will treat any SIP response to OPTIONS as an indication that the session agent is inservice. Therefore, the primary site Session Director will consider the session agent to the S8730
Processor Ethernet in-service, even though the S8500 ESS is in control of both sites. A sample
command output in this state is shown below:
09:06:17-35 (recent)
Active Rate ConEx
10.3.3.40
I
0
0.0
0
2.2.85.2
I
0
0.0
0
2.2.85.20
I
0
0.0
0
2.2.87.13
I
0
0.0
0

Active Rate ConEx
Avg
Max Burst
0
0.0
0 0.006 0.007
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.118 0.118
1
0
0.0
0 0.002 0.002
1
If an incoming PSTN call comes to the primary site Session Director, and the Session Director
chooses the session agent for the S8730 Processor Ethernet, the Session Director will receive a
503 Service Unavailable as a response to the inbound INVITE. The Session Director will then
automatically choose an alternate session agent in the same SAG to handle the call. This
behavior can be observed in the following Wireshark trace. In frame 38, the primary site Session
Director sends SIP OPTIONS to the S8730 Processor Ethernet. In frame 40, the S8730 responds
with 503 Service Unavailable. Absent changes to the default configuration, the 503 is a
sufficient response to mark the session agent in-service. In frame 46, the Service Provider
JRR; Reviewed:
SPOC 7/21/2009

49 of 70
CM-PE-NN4500
delivers an inbound call to the primary site Session Director. In frame 48, the INVITE is sent
from the Session Director to the S8730 Processor Ethernet. In frame 49, a 503 Service
Unavailable is returned to the Session Director. In frame 51, the Session Director delivers the
INVITE to 2.2.85.2, another session agent in the same SAG, and the call succeeds.
Since calls succeed in this state due to SAG recursion, the default behavior of the Session
Director for the session agent to the S8730 Processor Ethernet may be sufficient. If desired, the
acceptable SIP responses (e.g., 200 OK only) to SIP OPTIONS that mark a session agent inservice can be modified for the S8730 Processor Ethernet session agent, as discussed as in
Section 4.7.
5.3.2. Restoration of System Control back to the S8730 Servers

Now assume that the problem that caused the S8730 Servers to be off-line is resolved, and
system control is brought back to the active S8730 Server. For a short period of time, this
creates a scenario similar to the one described in Section 5.3.1. In this case, the ESS that had
recently been active will continue to respond to SIP OPTIONS from the secondary site Session
Director with 503 Service Unavailable for a brief period. By default, the Session Director will
treat any SIP response to OPTIONS as an indication that the session agent is in-service.
Therefore, the secondary site Session Director will consider the session agent to the ESS inservice, even though the ESS is no longer in control of resources.
If an incoming PSTN call comes to the secondary site Session Director, and the Session Director
chooses the session agent for the ESS Processor Ethernet, the Session Director will receive a 503
Service Unavailable as a response to the inbound INVITE. The Session Director will then
automatically choose an alternate session agent in the same SAG to handle the call.
This behavior can be observed in the following series of Wireshark traces.
JRR; Reviewed:
SPOC 7/21/2009

50 of 70
CM-PE-NN4500
The first trace below is filtered on SIP messages involving the ESS Processor Ethernet. In frame
38, the secondary site Session Director sends SIP OPTIONS to the ESS. In frame 39, the ESS
responds with 503 Service Unavailable, which absent changes to the default configuration, is a
sufficient response to keep the session agent in-service. This exchange re-occurs every 16
seconds in frames 156-157 and 1110-1111. In frame 2048, the Session Director sends an
INVITE to the ESS Processor Ethernet. In frame 2049, a 503 Service Unavailable is returned to
the Session Director. Since this trace only shows messages that involved the ESS Processor
Ethernet, the messages involving other session agents are not shown.
The screen below is from the same Wireshark trace, but with a different filter, to expose the
completion of the call. In frame 2046, the Service Provider sends the INVITE to the secondary
site Session Director. As stated above, the ESS responds with a 503, and in frame 2051, the
Session Director delivers the INVITE to another session agent in the same SAG, in this case
2.2.185.2, corresponding to a C-LAN now controlled by the active S8730 Server, which has
recovered system control. The remainder of the trace shows that the call proceeds normally.
Below, a later view into the same Wireshark trace is filtered on TCP and the ESS Processor
Ethernet, showing the ESS rejecting communications at the TCP level. As a result, the
secondary site Session Director will receive no SIP-level response from the ESS, and will mark
the session agent to the ESS out-of-service.
JRR; Reviewed:
SPOC 7/21/2009

51 of 70
CM-PE-NN4500
The condition where the ESS that had been actively controlling resources is no longer controlling
resources, but is still responding at the SIP stack level, is a transient condition. Since calls
succeed in this state due to SAG recursion, the default behavior of the Session Director for the
session agent to the ESS may be sufficient. If desired, the acceptable SIP responses (e.g., 200
OK only) to SIP OPTIONS that mark a session agent in-service can be modified, as discussed as
in Section 4.7.
5.4. Busy / Unavailable Conditions for Avaya Signaling / Trunk groups

If an Avaya trunk group is busied out (maintenance busy) or simply has no available trunk
members when an INVITE arrives for the corresponding signaling group, Communication
Manager will respond with a 503. This response will immediately cause the Session Director to
attempt to deliver the call to a different session agent in the SAG.
If an Avaya C-LAN based signaling group is busied out, the Session Director will mark the
session agent to the C-LAN out-of-service. An Avaya SIP signaling group using a C-LAN as the
near-end will not respond to SIP messages from the Session Director, but will instead reject
communications at the TCP level. The following portion of a Wireshark trace shows such a
condition. At the start of the trace, signaling group 30 is operating normally, and the C-LAN at
2.2.85.2 is responding to OPTIONS with 200 OK, as in frame 6. However, later in the trace,
after signaling group 30 has been busied out, TCP level rejections can be observed. The Session
Director tries to open a new TCP connection for transmitting OPTIONS, but will not receive a
SIP level response from a busied-out C-LAN based signaling group. Therefore, the Session
Director will mark the session agent out-of-service, and new calls will use other members of the
SAG, without trying the out-of-service session agent.
JRR; Reviewed:
SPOC 7/21/2009

52 of 70
CM-PE-NN4500
In contrast, if an Avaya S8730 Server Processor Ethernet SIP signaling group is busied out, the
Session Director (with default behavior) will retain an in-service state for the session agent. The
following portion of a Wireshark trace shows such a condition. At the start of the trace,
signaling group 32 is operating normally, and the active S8730 Server at 2.2.87.13 is responding
to OPTIONS with 200 OK, as in frame 213. However, later in the trace, after signaling group 32
has been busied out, the active S8730 Server responds with a 503 Service Unavailable, as shown
in frame 292 and 927. The Session Director will mark the session agent in-service (unless the
Session Director is configured to consider only a 200 OK acceptable via ping-in-serviceresponse-codes, see Section 4.7).
New calls can still succeed to other members of the SAG, but the call will be offered to the
busied-out Processor Ethernet signaling group, as shown in the trace below. In frame 1051, the
SIP Service Provider sends the incoming call INVITE to the primary site Session Director. In
frame 1053, the Session Director sends the INVITE to the S8730 session agent, since it is still
considered to be in-service. In frame 1054, Communication Manager responds with a 503. In
frame 1056, the Session Director sends the INVITE to another member of the same SAG, and
the call can complete normally from that point forward.
A Processor Ethernet based SIP signaling group will be marked out-of-service by the Session
Director, if the change ip-interface procr command is used to disable the ip-interface.
(However, note that doing so can affect other devices and applications). An example of this
condition is shown below. When the ip-interface procr is disabled, signaling groups with nearend procr will not respond at the SIP level.
See Section 5.4 of reference [JRR] for additional failure scenarios involving session agents.
In general, if there is one failed session agent at the time when an incoming call arrives, the
behavior depends on whether the Acme Packet Net-Net 4500 has realized the session agent is no
longer available. If the Session Director has not made the determination that the session agent is
unreachable, the session agent could be tried, subject to the SAG group strategy. The session
will experience a transaction timeout, and the session agent will be marked out-of-service.
JRR; Reviewed:
SPOC 7/21/2009

53 of 70
CM-PE-NN4500
The INVITE will be directed to a working session agent in the SAG, and the call will complete,
albeit with a slight delay (i.e., the transaction timeout).
Once the failed session agent has been marked out-of-service due to the transaction timeout, new
incoming calls will immediately be directed to the working session agent in the SAG, and the
call will complete without the delay. Note that a 503 response is not a transaction timeout and
therefore the session agent is not taken out-of-service by the Session Director.
As described in Section 1.3, Communication Manager can mark a SIP Signaling Group for
bypass and the corresponding SIP trunk members Out-of-Service/Far-end (OOS/FE) due to
failure of the SIP OPTIONS exchange. If the network recovers, but a successful SIP OPTIONS
exchange has not yet occurred, the Avaya trunk members may be OOS/FE when an incoming
INVITE arrives. Communication Manager will accept the incoming call. The following screen
shows an example where an incoming call was received by the S8730 Processor Ethernet
signaling group, just after recovery of the service provider network after an extended outage.
Prior to this call, all members of trunk group 32 were OOS/FE. After this call, all members of
trunk group 32 were marked in-service.
list trace tac 132
Page
LIST TRACE
time
14:13:37
14:13:37
14:13:37
14:13:37
14:13:37
14:13:37
14:13:37
14:13:41
14:13:41
14:13:41
data
Calling party trunk-group 32 member 1 cid 0x111a
active trunk-group 32 member 1 cid 0x111a
dial 52020
ring station
52020 cid 0x111a
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.26.4]:3488
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49196
rgn:3 [2.2.26.4]:3480
active station
52020 cid 0x111a
G711MU ss:off ps:20
rgn:3 [2.2.85.45]:49196
rgn:3 [2.2.1.109]:19036
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.85.45]:49196
If all session agents that are members of a SAG are not responding, or all trunks at a site are
busy, then the Acme Packet Net-Net 4500 will in general return a SIP 503 to the network.
Assuming the SIP Service Provider can redirect calls to the opposite site upon receiving a 503
from the initial site used for the call, then the call can complete successfully at the working site.
JRR; Reviewed:
SPOC 7/21/2009

54 of 70
CM-PE-NN4500
The following is a Wireshark trace for an incoming call when there is an all trunks busy
condition at the primary site. The trace is filtered to show only the inside communications
between the primary site Session Director and Communication Manager. At the start of the
trace, OPTIONS messages can be observed, with 503 responses, as is typical when there are no
trunks available to take a call. Using the default session agent behavior for SIP OPTIONS, the
session agents remain in-service. In frame 90, the Session Director sends an INVITE to 2.2.85.2,
the session agent corresponding to Avaya signaling group 30. In frame 94, Communication
Manager sends a 503. There are no available trunk members in trunk group 30 to handle the call.
In frame 96, the Session Director sends the INVITE to 2.2.85.20, another session agent in the
same SAG. In frame 101, Communication Manager sends a 503 because there are no available
trunk members in trunk group 31 either. In frame 103, the Session Director sends the INVITE to
2.2.87.13, the S8730 Processor Ethernet session agent in the same SAG. In frame 104,
Communication Manager sends a 503 because there are no available trunk members in trunk
group 32 either. Although not shown, the Session Director returns a 503 to the SIP Service
Provider. In this case, for calls to fail-over to an alternate site, the SIP Service Provider must
have the capability to redirect the call upon receipt of a 503 Service Unavailable from the
enterprise site.
5.5. Acme Packet Net-Net Session Director Off-line

This section pertains to cases where an Acme Packet Net-Net Session Director is off-line. For
example, during testing the Acme Packet Net-Net Session Director was powered down. Refer
to induced failure reference numbers 2 and 5 in Figure 1. Again, note that either or both sites
could use a pair of Acme Packet Net-Net Session Directors in a High Availability (HA)
configuration. The Acme Packet HA configuration was documented and tested in reference
[AC-HA].
5.5.1. Incoming Call from PSTN, Sites Acme Packet Net-Net 4500 Off-line
When the Acme Packet Net-Net 4500 is not responding at one site, it is expected that the SIP
Service Provider would redirect the call to the other site. Therefore, incoming calls can still
succeed, subject to the capabilities of the SIP Service Provider to provide fail-over. From a
Communication Manager point of view, incoming calls can arrive from either the primary site or
secondary site and reach any user. Therefore, inbound calls with one Acme Packet Net-Net 4500
offline look to Communication Manager like the call traces in Section 5.1, arriving from the
working Acme Packet Net-Net 4500.
5.5.2. Outgoing Call to PSTN, Sites Acme Packet Net-Net 4500 Off-line
The sample trace that follows was taken immediately after Ethernet was removed from the inside
interface of the Acme Packet Net-Net 4500 at the primary site. Extension 52020 dials the ARS
JRR; Reviewed:
SPOC 7/21/2009

55 of 70
CM-PE-NN4500
access code followed by 17328522550. The call is delivered to route pattern 30, which at the
time of this call, lists two trunk groups at the primary site (32, 31) followed by two trunks at the
secondary site (60, 61). Six seconds after trying trunk group 32, the denial event marked in bold
in the trace triggers LAR to the next choice in the pattern, trunk group 31. Since trunk group 31
is also terminating on the off-line primary site Session Director, the bold denial event triggers
LAR to the next choice, trunk group 60. Since trunk group 60 is operating normally using the
secondary site Session Director, the call completes using trunk group 60. Note that the six
seconds is governed by the timer named Alternate Route Timer on the Avaya signaling group.
Six seconds is the default value. Ultimately, the final connection is an inter-region ip-direct
connection between the IP Telephone (2.2.1.109) at the primary site and the Acme Packet NetNet 4500 (2.2.185.145) at the secondary site.
Page
LIST TRACE
time
14:21:22
14:21:22
14:21:25
14:21:25
14:21:34
14:21:34
14:21:34
14:21:34
14:21:34
14:21:40
14:21:40
14:21:40
14:21:40
14:21:40
14:21:40
14:21:40
14:21:40
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:46
14:21:52
14:21:52
14:21:52
data
active station
52020 cid 0x111c
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.26.4]:3512
term trunk-group 32
cid 0x111c
route-pattern 30 preference 1 cid 0x111c
seize trunk-group 32 member 2 cid 0x111c
denial event 1191: Network failure D1=0x8c51 D2=0x26
route-pattern 30 preference 1 unavailable cid 0x111c
term trunk-group 31
cid 0x111c
term trunk-group 32
cid 0x111c
route-pattern 30 preference 2 unavailable cid 0x111c
term trunk-group 60
cid 0x111c
Proceed trunk-group 60 member 2 cid 0x111c
G711MU ss:off ps:20
rgn:1 [2.2.185.145]:49152
rgn:3 [2.2.26.4]:3560
active trunk-group 60 member 2 cid 0x111c
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:1 [2.2.185.145]:49152
G711MU ss:off ps:20
rgn:1 [2.2.185.145]:49152
rgn:3 [2.2.1.109]:19036
JRR; Reviewed:
SPOC 7/21/2009

56 of 70
CM-PE-NN4500
If the failure persists, Communication Manager will mark the signaling group for bypass (and
trunk groups OOS/FE). Therefore, outbound calls like the one traced above would immediately
proceed to the trunks connecting to the working Acme Packet Net-Net 4500, without the delay
associated with the timeout of the outbound INVITEs. LAR is only required when the trunks are
chosen because they appear to be in-service, and then a timeout or down-stream failure occurs
requiring a route-advance.
If an Avaya SIP trunk group is in the Out-of-Service/Far-end state (OOS/FE), Communication
Manager will accept an incoming call (i.e., accept INVITE), but will not offer an outgoing call
(i.e., send INVITE). In the sample configuration, if the Acme Packet Net-Net 4500 at the farend of the Avaya signaling group has a failure that persists, this state will be seen. This state
will also be seen if the Acme Packet Net-Net 4500 is functioning properly, but the SIP Service
Provider that is the next hop has failed. The Acme Packet Net-Net 4500 will respond with a
503 when the next hop is out of service.
Failure of the secondary site Acme Packet Net-Net 4500 produces similar results.
5.6. SIP Service Provider Network Off-line

This section pertains to cases where all enterprise components are functioning, but the SIP
Service Provider network is not. Refer to induced failure reference numbers 7 and 8 in Figure 1.
5.6.1. Outbound Calls to PSTN

The Communication Manager route-pattern configuration can contain trunks of various kinds. In
the sample configuration, only SIP trunk groups are shown in route pattern 30 and 60. In
production environments, it will be common to have traditional trunks such as ISDN-PRI trunks
available as an alternate to the SIP trunks A call to a route pattern that preferentially uses SIP
trunks can overflow or look-ahead and complete successfully to traditional trunks later in the
pattern.
JRR; Reviewed:
SPOC 7/21/2009

57 of 70
CM-PE-NN4500
The following screen shows an example call trace when a user at the primary site attempts an
outbound PSTN call right after the failure of the service provider network. That is,
Communication Manager can still communicate with the Acme Packet Net-Net 4500, but the
Acme Packet Net-Net 4500 at the primary site is unable to communicate with the public
network, and the Acme Packet Net-Net 4500 has not yet marked the session agent to the public
network out-of-service. Since Communication Manager has not yet marked the trunks OOS, the
call is offered to the first choice in the route-pattern (a primary site Processor Ethernet trunk),
then the second choice in the route-pattern (a primary site C-LAN trunk), then the third choice, a
secondary site trunk, and so on.
Page
LIST TRACE
time
15:09:34
15:09:34
data
active station
52020 cid 0x1122
G711MU ss:off ps:20
rgn:3 [2.2.1.109]:19036
rgn:3 [2.2.26.4]:3836
15:09:38
15:09:38
term trunk-group 32
cid 0x1122
15:09:44
15:09:44
15:09:44
15:09:44
15:09:44
15:09:44
15:09:50
15:09:50
route-pattern 30 preference 1 unavailable cid 0x1122
15:09:50
15:09:50
term trunk-group 31
cid 0x1122
15:09:50
15:09:50
15:09:50
15:09:50
15:09:50
15:09:54
15:09:54
term station
52020 cid 0x1122
15:09:56
15:09:56
term trunk-group 32
cid 0x1122
15:09:56
route-pattern 30 preference 2 unavailable cid 0x1122
15:09:56
15:09:56
term trunk-group 60
cid 0x1122
15:09:56
15:09:56
15:09:56
<and so on, text removed>
In the case shown above, the Wireshark trace would show that the Acme Packet Net-Net 4500 at
the primary site responds with 100 TRYING to INVITE messages, but Communication Manager
would receive nothing after 100 TRYING. Communication Manager does a route-advance
due to LAR = next on the route-pattern six seconds later.
Now assume the network outage persists long enough for the Acme Packet Net-Net 4500 at the
primary site to mark the session agent to the public network out-of-service. However, the outage
has not persisted long enough for Communication Manager to have marked the trunks to the
Session Director out-of-service. In the sample configuration, the Session Director sources SIP
JRR; Reviewed:
SPOC 7/21/2009

58 of 70
CM-PE-NN4500
OPTIONS every 16 seconds, so the Session Director will generally discover that the session
agent to the service provider is out before Communication Manager. In this case, since
Communication Manager has not yet marked the trunks OOS, the call is offered to the first
choice in the route-pattern (a primary site trunk). Since the Session Director has already marked
the next-hop out-of-service, the Session Director responds immediately with a 503.
Communication Manager can therefore immediately route-advance (no need to wait for timeout)
to the next choice in the route-pattern. In other words, once the Session Director has detected
that the Service Provider is non-responsive, Look-ahead Routing will occur quickly. An
example trace is provided in Section 5.6 of reference [JRR].
Once Communication Manager determines via SIP OPTIONS testing that the network is out (see
Section 1.3 and Section 5.7 of reference [JRR] for details), a similar call would simply bypass
the trunks that are marked out-of-service.
5.7. Wireshark Traces Illustrating SIP OPTIONS Behaviors

See Section 1.3 and Section 5.7 of reference [JRR] for additional information, including SIP
OPTIONS exchanges related to the session agents and signaling groups for the C-LAN SIP
interfaces. The sub-sections below highlight behaviors associated with the Processor Ethernet
signaling groups and session agents.
5.7.1. OPTIONS Sourced by Communication Manager

The following portion of a wireshark trace illustrates a SIP OPTIONS exchange initiated by
Communication Manager for signaling group 32, when the network is functioning normally.
Signaling group 32 has near-end IP 2.2.87.13, the procr of the active Avaya S8730 Server.
The far-end of signaling group 32 is 2.2.85.45, the inside address of the Acme Packet Net-Net
4500 at the primary site. In frame 16, Communication Manager sends the SIP OPTIONS to the
Session Director. In frame 17, the Session Director forwards the SIP OPTIONS to the outside
SIP Service Provider network, session agent 10.3.3.40. In frame 18, the network responds with
200 OK. In frame 19, the Session Director responds to Communication Manager with 200 OK.
The Avaya signaling group is in-service. Note that Wireshark was configured such that TCP
messages would appear in gray, and UDP messages in blue.
The following portion of a Wireshark trace illustrates a SIP OPTIONS exchange initiated by
Communication Manager for signaling group 62, also when the enterprise network is functioning
normally. Signaling group 62 has near-end procr, which is the IP Address of the active Avaya
S8730 Server in this case. The far-end of signaling group 62 is 2.2.185.145, the inside address of
the Acme Packet Net-Net 4500 at the secondary site. In frame 117, Communication Manager
opens a TCP connection to the Session Director at the secondary site. In frame 120,
Communication Manager sends the SIP OPTIONS to the Session Director for signaling group
62. In frame 121, the Session Director responds with a TCP FIN, ACK. The active Avaya
S8730 Server will not receive a 200 OK response from the Session Director at the secondary site,
because the secondary site does not have a session agent defined for the Avaya S8730 Server at
JRR; Reviewed:
SPOC 7/21/2009

59 of 70
CM-PE-NN4500
the primary site. Communication Manager marks signaling group 62 for bypass and the
corresponding trunk group 62 Out-of-Service/Far-end.
Communication Manager will periodically try to establish the connectivity for signaling group
62, each time with a similar exchange (using a new source port) as shown below. Signaling
group 62 will always be out-of-service from the point of view of Communication Manager
running on the Avaya S8730 Server.
5.7.1.1 Communication Manager running load 947.3

The traces in this section were taken while Communication Manager load 947.3 was running
(i.e., without SP1).
When the ESS is active (e.g., network failure between the primary site and the secondary site),
Communication Manager running on the ESS can successfully exchange SIP OPTIONS with the
Session Director at the secondary site, bringing signaling group and trunk group 62 in-service on
the ESS. With reference to the trace below, in frame 3356, Communication Manager on the
active ESS sends SIP OPTIONS to the Session Director at the secondary site. In frame 3357, the
Session Director forwards the SIP OPTIONS to the outside SIP Service Provider network,
session agent 10.3.3.1. In frame 3358, the network responds with 200 OK. In frame 3359, the
Session Director responds to Communication Manager with 200 OK. Avaya signaling group 62
is in-service on the ESS, and can remain so.
JRR; Reviewed:
SPOC 7/21/2009

60 of 70
CM-PE-NN4500
When the ESS is active, Communication Manager running on the ESS will also attempt to bring
up signaling group 32. However, signaling group 32 will always remain out-of-service from the
point of view of the ESS for the same reason that signaling group 62 will always remain out-ofservice from the point of view of the Avaya S8730. In frame 3406, the ESS opens a TCP
connection to the primary site Session Director. (At the time of this trace, the WAN failure
causing the ESS to becoming active had been resolved). In frame 3409, the ESS sends
OPTIONS to the Session Director. Communication Manager on the ESS may receive responses
as shown below at the TCP level, but will not receive a 200 OK from the primary site Session
Director, since the primary site Session Director does not have a configured session agent for the
ESS. The ESS continues to try (new TCP source port) beginning in frame 3566.
5.7.1.2 Communication Manager running load 947.3 plus SP1

The traces in this section were taken while Communication Manager load 947.3 plus SP1 was
running. A change introduced in SP1 causes an active ESS to use the ping method to maintain
a signaling group, even if the Enable Layer 3 Test parameter is set to y on the signaling
group. See Section 6 for additional information on the modification request associated with this
observation.
In the trace below, frames 748 through 750 show SIP OPTIONS sourced by the secondary site
Session Director, and the corresponding 200 OK from the active ESS. The frames that follow
show that Communication Manager on the active ESS sends ICMP ping requests to the
secondary site Session Director, rather than SIP OPTIONS. In this case, the Session Director
was configured not to respond to ping on any interface other than the management interface.
Therefore, Communication Manager does not get a response. As a result, signaling group 62 is
marked for bypass and trunk group 62 is marked out-of-service/far-end. A work-around is
documented in Section 4.3.
The trace below shows that the active ESS is also trying to maintain signaling group 32 to the
primary site Session Director with ICMP ping. Again, since the Session Director was configured
not to respond to ping on any interface other than the management interface, there is no response.
Signaling group 32 is therefore marked for bypass like signaling group 62. This trace is
included to emphasize that if the work-around to enable ping responses is implemented on the
JRR; Reviewed:
SPOC 7/21/2009

61 of 70
CM-PE-NN4500
Session Director, it should be implemented on the secondary site Session Director only, to
prevent signaling group 32 from coming in-service on the active ESS when the primary site
Session Director has no session agent to handle SIP requests from the ESS at the secondary site.
The following Wireshark trace shows the behavior after the work-around documented in Section
4.3 is implemented for the secondary site Session Director. Frames 729-837 show the Session
Director maintaining the session agent to the active ESS via SIP OPTIONS. In frame 852, the
ESS Processor Ethernet sources an ICMP ping request to the Session Director. Since the workaround that allows the inside interface of the secondary site Session Director to respond to ping
has been implemented, frame 853 shows the ping reply. Several other request / reply pairs are
also evident. Signaling group 62 remains in-service.
5.7.2. OPTIONS Sourced by Acme Packet Session Director

The following portion of a Wireshark trace shows the Session Director at the primary site
(2.2.85.45) maintaining the session agent to the active S8730 Server at 2.2.87.13, under normal
conditions. In frame 58, the Session Director at the primary site sends the SIP OPTIONS, and
in frame 60, Communication Manager on the S8730 Server responds with 200 OK. This repeats
every 16 seconds (i.e., the configured interval), absent other traffic, as shown below.
The following portion of a Wireshark trace shows the Session Director at the secondary site
(2.2.185.145) attempting to bring up the session agent to the ESS at 2.2.185.88, under normal
conditions, when the ESS is not active. In frame 328, the Session Director at the secondary site
JRR; Reviewed:
SPOC 7/21/2009

62 of 70
CM-PE-NN4500
opens a TCP connection to the ESS. In frame 331, the Session Director sends the SIP
OPTIONS. In frame 334, the ESS responds with a TCP FIN, ACK. The ESS will not respond at
the SIP level when the ESS is not active.
The Session Director will continue to try to establish the connectivity for the session agent
corresponding to the ESS procr, but this session agent will remain out-of-service until the ESS
is activated (e.g., by a network or other failure). The following portion of a Wireshark trace
illustrates another attempt (source TCP port will count up by 2 for each attempt)
The following portion of a Wireshark trace shows the Session Director at the secondary site
successfully maintaining the session agent to the ESS in-service, when the ESS is active (in this
case, controlling cabinet 1). Frame 76 shows Communication Manager running on the ESS
responding with a 200 OK. Recall that the Session Director was configured to send SIP
OPTIONS every 16 seconds in the absence of other traffic. The 16 second interval between SIP
OPTIONS is evident in the trace below (second column timestamp).
5.8. S8730 Server Interchange

This section illustrates the effect on SIP signaling to the primary site Session Director upon an
interchange of the S8730 Servers. During an interchange, the active S8730 server becomes the
standby, and the previous standby server becomes the active. Since the primary site Session
Director has a session agent to the active server IP (2.2.87.13), the physical server that the
Session Director is communicating with for this session agent changes after the interchange. In
the sample configuration, the inside interface of the Session Director and the S8730 Servers are
in different subnets, as would be typical in production deployments. The interchange becomes
evident at the TCP level, as evidenced in the following portions of the same Wireshark trace.
JRR; Reviewed:
SPOC 7/21/2009

63 of 70
CM-PE-NN4500
The following portion of a Wireshark trace, filtered for TCP, shows the TCP connections in-use
prior to an interchange. In this case, when the Session Director sends SIP OPTIONS to the
S8730 Server session agent at 2.2.87.13, it uses TCP source port 8196 and destination port 5060,
as can be seen in frames 508-571 below. When Communication Manager sends SIP OPTIONS
to maintain signaling group 32 using the S8730 Processor Ethernet, it opens a TCP connection
from source port 30166 to destination port 5060, sends the SIP OPTIONS, and receives a 200
OK, as can be seen in frames 603-610. (These source ports will vary).
Subsequent communications will use these TCP connections. For example, the following
portion of the same trace shows an incoming call from the PSTN to an Avaya IP Telephone. In
frame 2171, the Session Director delivers the INVITE. The ACKs in frames 2176, 2187, and
3611 show that port 8196 is used for the call activity, the same port used for Session Director
sourced OPTIONS in the prior trace. After the call is answered, Communication Manager
begins the process of shuffling the call to ip-direct media paths, with the INVITE in frame
3613. This communication used source port 30166 (see frame 3622), and this is also the same
port used for Communication Manager outbound OPTIONS in the prior trace.
JRR; Reviewed:
SPOC 7/21/2009

64 of 70
CM-PE-NN4500
With this call up, an S8730 Server interchange was requested via the Communication Manager
maintenance Web page. The following portion of the same trace shows the activity before,
during, and after the interchange. In frames 29500 through 32193, the normal Acme Packet
OPTIONS messaging can be seen, still using source port 8196. In frames 35228 and 35229,
Communication Manager sends a TCP FIN, ACK for the two connections that were established.
In frame 36133, Communication Manager (running on the newly active S8730 Server) begins the
process of establishing a new TCP connection to the Session Director, from source port 30383 to
destination port 5060. In frame 36136, this new connection is used to send SIP OPTIONS from
Communication Manager to the Session Director. In frame 36267, the Session Director begins
the process of opening a new TCP connection from source port 8198 to 2.2.87.13 at port 5060
(i.e., to the newly active S8730 Server). In frame 36270, the Session Director sends SIP
OPTIONS using the new connection.
At this point, subsequent SIP signaling will use the two new TCP connections. For example, the
following portion of the same trace shows a successful Hold / Resume sequence for the call that
was established prior to the server interchange. The INVITE in frame 42418 initiates the Hold
messaging, using source port 30383, the same port Communication Manager used to send
OPTIONS immediately after the interchange. The INVITE in frame 43015 initiates the Resume
(un-hold) messaging. When the Session Director sends OPTIONS (frame 43447), it continues to
use port 8198.
JRR; Reviewed:
SPOC 7/21/2009

65 of 70
CM-PE-NN4500
The next screen shows a portion of the same Wireshark trace during the interchange, but using a
different filter, to show TCP communications between a C-LAN session agent (2.2.85.2) and the
Session Director. As seen previously, the interchange took place around frame 35228 and time
stamp 374. The trace below includes messages before and after the interchange. Observe that all
SIP OPTIONS sourced by the Session Director to the C-LAN use source port 8192, which
Wireshark has decoded to spytechphone. This screen is included to contrast the C-LAN based
session agent, for which no new TCP connection was needed as a result of the S8730
interchange, with the S8730 Server session agent, which requires a new TCP connection to be
established to the newly active S8730 Server after the interchange.
The next portion of a new Wireshark trace shows a normal call hang-up by the originating PSTN
user for the call highlighted in this section. Like the OPTIONS messages after the interchange,
the BYE in frame 6481 uses source port 8198. The trace concludes with SIP OPTIONS from
Communication Manager, continuing to use source port 30383.
6. Test Observations Leading to Product Modification

Requests
This section documents observations made during testing that stimulated product modification
requests.
Avaya Aura Communication Manager Release 5.2 is the first Communication Manager release
to include the P-Charging-Vector for outbound SIP trunk calls. A modification request has been
entered to change the way that the ICID in the P-Charging-Vector is encoded. Until the
modification request is resolved, the Acme Packet Net-Net Session Director configuration can
delete the P-Charging-Vector to avoid revealing private side IP Address information to the
public network. Section 4.6 of reference [JRR] documents the relevant configuration. The
JRR; Reviewed:
SPOC 7/21/2009

66 of 70
CM-PE-NN4500
product changes associated with the Avaya modification request were included in Service Pack 1
for Release 5.2. During the testing associated with these Application Notes, it was confirmed
that with Service Pack 1, Communication Manager no longer reveals the private side IP Address
in the ICID of the P-Charging-Vector. That is, using Service Pack 1, it is not necessary for the
Session Director to be configured to delete the P-Charging-Vector.
Section 5.2.4 covers outbound calls made by users at the ESS site, when the system is
fragmented, such that the ESS controls the secondary site, and the active S8730 Server controls
the primary site. In the sample configuration, PSTN calls from secondary site users are directed
to a route pattern that contains local secondary site SIP trunks before primary site SIP trunks.
Therefore, outbound calls from users at the secondary site will succeed, using the SIP trunks at
the secondary site. In an alternate configuration, where calls from secondary site users
controlled by the ESS could be directed to route patterns that contain SIP trunks at the primary
site before SIP trunks at the secondary site, outbound calls can potentially fail. A modification
request was entered, and a fix was delivered to Service Pack 1 for Release 5.2. During the
testing associated with these Application Notes, it was confirmed that with Service Pack 1, this
problem is resolved.
With the generally available Acme Packet Session Director release used for the testing, the
Session Director could respond with a SIP 503 to Communication Manager sourced OPTIONS
messages after certain types of failures. For example, if the Ethernet connectivity to a C-LAN
were removed for a minute, and then re-inserted, Communication Manager would send
OPTIONS to the Acme Packet Net-Net 4500 as part of restoring the SIP signaling group. These
initial SIP OPTIONS from Communication Manager resulted in a 503 response. Despite the
503, the Session Director did mark the session agent in-service, and could deliver calls to
Communication Manager using the restored C-LAN session agent. However, in the absence of
incoming call activity, Communication Manager would consider the corresponding trunks Outof-service/Far-end until an OPTIONS exchange sourced by Communication Manager succeeds.
This generally occurred in approximately 5-7 minutes. An Acme Packet ticket, 18281, was
entered to document the problem. A workspace with a fix was also delivered to Avaya for
testing, which was tested and did resolve the problem. Although not verified, Acme Packet later
informed Avaya that the fix was delivered to generally available release SCX6.1.0 MR-1 Patch 2
(nnSCX610m1p2) for the Net-Net 4500 Session Director.
As described in Section 5.7, Communication Manager SP1 changed the behavior of maintenance
associated with SIP signaling groups, when the ESS is activated. Prior to SP1, SIP signaling
groups controlled by an ESS followed the configuration of the Enable Layer 3 Test field on the
signaling group. That is, if y, OPTIONS messages were used rather than ICMP pings to check
the availability of the far-end of the signaling group. Using SP1, signaling groups controlled by
an ESS use the ICMP ping approach, regardless of the configuration of the Enable Layer 3
Test parameter on the signaling group. A modification request was entered. At time of writing,
reversion to the prior behavior, where the Enable Layer 3 Test configuration is honored by the
active ESS, is targeted for Communication Manager 5.2 SP3 (and Communication Manager
5.2.1). Until that time, a work-around is documented in Section 4.3. Recall that inbound calls
are accepted on a trunk in the OOS/FE state (signaling group in bypass), and an inbound call
JRR; Reviewed:
SPOC 7/21/2009

67 of 70
CM-PE-NN4500
will cause all trunk members to move to an in-service state. A deployment with inbound calls
would keep the trunks controlled by the ESS in-service, even without the work-around.
7. Conclusion
As illustrated in these Application Notes, Communication Manager 5.2 can use the Processor
Ethernet of both an S8730 Server and an ESS as the signaling interface for SIP signaling groups
and trunk groups. Like the C-LAN based SIP trunks documented in reference [JRR], the
Processor Ethernet based SIP trunks can interoperate with Acme Packet Net-Net 4500 as part of
a survivable SIP Trunking solution.
8. Additional References
This section references the documentation relevant to these Application Notes. Avaya product
documentation is available at http://support.avaya.com. Acme Packet product documentation is
available at http://www.acmepacket.com. A support account may be required to access the
Acme Packet documentation.
Reference [JRR] describes the base configuration and verification results for the network used as
the starting point for the configuration illustrated in these Application Notes.
[JRR] Sample Configuration Illustrating Avaya Aura Communication Manager SIP Trunk
Survivability with Enterprise Survivable Server and Acme Packet Net-Net 4500 Session
Director, Issue 1.0
<Paste link here once the Avaya web site posts it>
[JSR] Application Notes for Configuring Direct SIP Trunking from Communication Manager
using an Acme Packet Net-Net Session Director and a SIP PSTN Gateway
http://www.avaya.com/master-usa/en-us/resource/assets/applicationnotes/cm4acmesippstn.pdf
The following Application Notes show two independent sites running Communication Manager
networked via SIP Trunks to a pair of Acme Packet Net-Net 4500 configured in a High
Availability Configuration:
[AC-HA] Application Notes for Configuring Acme Packet Net-Net 4500 Session Director with
Direct SIP Trunking to Avaya Communication Manager, Issue 1.0
http://www.avaya.com/master-usa/enus/resource/assets/applicationnotes/Acme4500CM5DTrk.pdf
[LAR] Sample Configuration for SIP Private Networking and SIP Look-Ahead Routing Using
Avaya Communication Manager, Issue 1.0
http://www.avaya.com/master-usa/en-us/resource/assets/applicationnotes/sip-pvt-lar.pdf
[DPT] Configuring Avaya Communication Manager for Dial-Plan Transparency and InterGateway Alternate Routing, Issue 1.1
http://www.avaya.com/master-usa/en-us/resource/assets/applicationnotes/dpt-igar.pdf
JRR; Reviewed:
SPOC 7/21/2009

68 of 70
CM-PE-NN4500
[CM1] Administering Avaya Aura Communication Manager, Document Number 03-300509,

Release 5.2, May 2009.
http://support.avaya.com/elmodocs2/comm_mgr/r5.0/03-300509_4.pdf
[CM2] Avaya Aura Communication Manager Feature Description and Implementation,
Document Number 555-245-205, Issue 7, May 2009
http://support.avaya.com/elmodocs2/comm_mgr/R5_2/555_245_205_7.pdf
[CM3] SIP Support in Avaya Aura Communication Manager Document Number 555-245206, Issue 9, May 2009
http://support.avaya.com/elmodocs2/comm_mgr/R5_2/555_245_206_9.pdf
[ESS] Using the Avaya Enterprise Survivable Servers, Document Number 03-300428, Issue 5,
May 2009
http://support.avaya.com/elmodocs2/comm_mgr/R5_2/03_300428_5.pdf
[AP1] Net-Net 4000 ACLI Configuration Guide Release S-C6.1.0 January 2009, Document
Number 400-0061-61 Rev 1.01
JRR; Reviewed:
SPOC 7/21/2009

69 of 70
CM-PE-NN4500
2009 Avaya Inc. All Rights Reserved.
Avaya and the Avaya Logo are trademarks of Avaya Inc. All trademarks identified by and
are registered trademarks or trademarks, respectively, of Avaya Inc. All other trademarks are the
property of their respective owners. The information provided in these Application Notes is
subject to change without notice. The configurations, technical data, and recommendations
provided in these Application Notes are believed to be accurate and dependable, but are
presented without express or implied warranty. Users are responsible for their application of any
products specified in these Application Notes.
Please e-mail any questions or comments pertaining to these Application Notes along with the
full title name and filename, located in the lower right corner, directly to the Avaya Solution &
Interoperability Test Lab at interoplabnotes@list.avaya.com.
JRR; Reviewed:
SPOC 7/21/2009

70 of 70
CM-PE-NN4500

Acme SBC Config

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Acme SBC Config

Uploaded by

Copyright:

Available Formats

Avaya Solution & Interoperability Test Lab

Sample Configuration Illustrating Avaya Aura

Solution & Interoperability Test Lab Application Notes

Figure 1: Avaya Aura Communication Manager Survivable SIP Trunking

Solution & Interoperability Test Lab Application Notes

Figure 2: Avaya Aura Communication Manager Processor Ethernet SIP Trunking

Solution & Interoperability Test Lab Application Notes

1.1. Summary of Inbound Calls to the Enterprise

Solution & Interoperability Test Lab Application Notes

Figure 3: Incoming Calls to the Enterprise from the PSTN

Solution & Interoperability Test Lab Application Notes

1.2. Session Director and Communication Manager Terminology

The session-agent defines a SIP peer, similar to

Solution & Interoperability Test Lab Application Notes

1.3. SIP OPTIONS Message, Service States, and Call Acceptance

Solution & Interoperability Test Lab Application Notes

1.4. Summary of Outbound Calls from the Enterprise to the PSTN

Solution & Interoperability Test Lab Application Notes

2. Equipment and Software Validated

Avaya 4600-Series Telephones (H.323)

Avaya 9600-Series Telephones (H.323)

Release 3.0 H.323

Avaya 2400-Series Digital Telephones

Cisco 3825 Integrated Services Router

Avaya S8730 Servers (Main)

Avaya S8500 Server (ESS)

Table 1: Equipment and Software Versions Used

3. Avaya Aura Communication Manager Configuration

3.1. Configuration Processor Ethernet Via the Web Pages

Solution & Interoperability Test Lab Application Notes

Solution & Interoperability Test Lab Application Notes

Solution & Interoperability Test Lab Application Notes

Solution & Interoperability Test Lab Application Notes

3.2. Node Names

3.3. Adding the IP Interface for procr

Allow H.323 Endpoints? y

Solution & Interoperability Test Lab Application Notes

-- ------ ----- -------y C-LAN 01A02 TN799 D

y MEDPRO 02A07 TN2602

3.4. SIP Signaling Groups Using Processor Ethernet

Solution & Interoperability Test Lab Application Notes

Group Type: sip

Far-end Node Name: nn4500-prisite

Far-end Domain: 2.2.85.45

Direct IP-IP Audio Connections?

Group Type: sip

Far-end Node Name: nn4500-secsite

Far-end Domain: 2.2.185.145

Direct IP-IP Audio Connections?

3.5. SIP Trunk Groups

Solution & Interoperability Test Lab Application Notes

Group Type: sip

Redirect On OPTIM Failure: 5000

Numbering Format: public

Solution & Interoperability Test Lab Application Notes

3.6. Route Patterns

Solution & Interoperability Test Lab Application Notes

BCC VALUE TSC CA-TSC

ITC BCIE Service/Feature PARM

No. Numbering LAR

BCC VALUE TSC CA-TSC

ITC BCIE Service/Feature PARM

No. Numbering LAR