You are on page 1of 67

CCNA

Table of Contents
Introduction 1
Advantages of OSI Layered Model 1
7 OSI Layers 1
WAN Data-Link Protocols 2
HDLC 2
SDLC 2
LAPB 2
X.25 2
SLIP 2
PPP 2
Frame-Relay 3
Ethernet & 802.3 3
CSMA/CD 3
CSMA/CD Steps for a NIC 3
Ethernet Broadcasts 3
FDDI 3
Token Ring 4
ATM 5
LAN Segmentation 5
LAN Switching Types 5
Switches 5
Switching Modes 5
Multilayer Switching 5
Layer 3 Switching 5
Layer 3 Switching Options 6
Layer 4 Switching 6
Cisco MLS (Multi-Layer Switching) 6
Spanning Tree Protocol (STP - 802.1d) 6
VLAN (Virtual LAN) 6
Frame Tagging 6
Half-Duplex Ethernet 7
Full-Duplex Ethernet 7
Crossover Ethernet Cabling 7
MDI/MDX Buttons 7
Fast Ethernet 7
DoD Reference Model 8
TCP/IP Protocol Suite 9
Telnet 9

Copyright © 2004 cisconotes i


FTP 9
TFTP 9
NFS 9
SMTP 9
X-Window 9
SNMP 9
Host-to-Host 9
TCP 10
UDP 10
Internet Layer Protocols 11
RARP (Reverse Address Resolution Protocol) 11
BootP (Bootstrap Protocol) 11
ICMP (Internet Control Message Protocol) 11
Network Access Layer Protocols 12
Frames 12
ARP and the ARP Cache 12
IP Addressing 13
Private Addresses 13
Subnetting a Network 13
Cisco Routers 14
Router Components 15
Startup Sequence 15
Exec Command Interpreter 15
Manual Configuration 15
Terminal Editing Commands 16
Router Commands 16
IOS v10.3 and Earlier Commands 17
Passwords 17
System Configuration Dialog 18
DTE/DCE Cable 18
Routing Basics 19
Benefits of Static Routes 19
Dynamic Routing 20
Routed vs. Routing Protocols 20
Interior Routing Protocols 21
Distance-Vector Routing Protocols 21
Balanced Hybrid 22
Routing Information Protocol (RIP) 22
Interior Gateway Routing Protocol (IGRP) 23
Enhanced Interior Gateway Routing Protocol (EIGRP) 24
Open Shortest Path First (OSPF) 24
Exterior Routing Protocols 25
Exterior Gateway Protocol (EGP) 25
Border Gateway Protocol (BGP) 25

Copyright © 2004 cisconotes ii


Configuring Routing 25
Configuring RIP 25
Monitoring RIP 26
Configuring IGRP 27
Monitoring IGRP 27
Router Configuration 27
IOS Sources 27
Fallback and Default Location of IOS 27
Cisco AutoInstall Procedure 28
IOS commands 28
Accessing Other Routers 30
Virtual Terminal Connections (Telnet) 30
Hostnames 31
Basic Testing 32
PING Results Defined 32
Testing with the Trace Command 32
Other Testing Commands 33
IPX/SPX Protocol Suite 34
IPX (Internetwork Packet eXchange) 34
SPX (Sequence Packet eXchange) 34
RIP (Routing Information Protocol) 34
SAP (Service Advertising Protocol) 34
NLSP (Netware Link Services Protocol) 34
NCP (NetWare Core Protocol) 34
Server-Sever Communication 35
IPX Addressing 35
Enabling IPX on Cisco Routers 36
Enabling IPX on Individual Interfaces 36
Adding Secondary Addresses 37
Monitoring and Troubleshooting IPX on Cisco Routers 37
Extended Ping 39
Access Lists & Traffic Management 39
Standard IP Access Lists 40
Wildcard Masking 40
Extended IP Access Lists 40
Monitoring Extended IP Access Lists 40
Standard IPX Access Lists 41
Extended IPX Access Lists 41
IPX SAP Filters 41
VLANs (Virtual LANs) 42

Copyright © 2004 cisconotes iii


VLAN Types 42
Broadcasts 42
Security 43
Scalability & Flexibility 43
Distribution of Traffic 43
Distribution of Network Services 43
VLAN Features 43
VLANs Over Several Backbones 44
Configuring a New VLAN 44
Static VLANs 44
Dynamic VLANs 45
VMPS (VLAN Management Policy Server) 45
Dynamic Port VLAN Membership 45
Menu-Based VLAN Configuration 45
VTP Advertisements 47
Catalyst 5000 VTP Configuration 47
VTP Pruning 48
Spanning Tree (STP) Configuration 50
STP Verification 51
Wide Area Networking 52
Synchronous Data-Link Control (SDLC) 53
High-Level Data-Link Control (HDLC) 53
Transfer Modes Supported by HDLC 54
Dial-on-Demand routing (DDR) 54
X.25 54
3 Categories of X.25 Network Devices 55
X.25 Sessions 55
Virtual Circuits Over an X.25 Network 55
X.25 Protocol Suite 55
Five modes of operation 56
Link Access Procedure Balanced (LAPB) 56
X.21bis 56
X.25 on Cisco Routers 57
Frame Relay 57
Frame Relay with Cisco Routers 58
DLCIs (Data-Link Connection Identifiers) 58
Local Management Interface (LMI) 58
Subinterfaces 58
Monitoring Frame Relay 59
Point-to-Point Protocol (PPP) 59
Configuration 59
ISDN (Integrated Services Digital Network) 60
ISDN Terminals 60
ISDN Reference Points 60
ITU-T Defined Protocols for Diverse Issues 60

Copyright © 2004 cisconotes iv


Basic Rate Interface 60
Primary Rate Interface 60
ISDN with Cisco Routers 61
Sample Show Configuration Command on a T1 61

Copyright © 2004 cisconotes v


Introduction
Advantages of OSI Layered Model
-Clarifies general function, not specifics
-Complex network into manageable layers
-Standard interfaces Æ interoperability
-Developers can change the features of just one layer
-Allows specialization
-Eases troubleshooting

7 OSI Layers
Application
-Network Applications: file, print, message, application, and database services
-examples: WWW, E-Mail gateways, EDI, SIG BBS, Gopher, WAIS, Yahoo, Financial applications such
as Oracle-based budgeting software, NT or Win2000 Server Applications.

Presentation (Translator)
-Compress/Decompress
-Encrypt/Decrypt
-SNMP uses ASN.1 Abstract Syntax Notation – standard data syntax used in this layer.
-PICT, TIFF, JPEG, MIDI, MPEG, WMA, QuickTime

Session (Dialog control between devices (nodes))


-Simplex, half-duplex, full-duplex (with flow control)
-Connection-oriented: formal session with check-points for large data
-3 phases: connection establishment – agree on communication parameters, data transfer, systematic
release
-Or connectionless
-NFS, SQL, RPC, X-Windows, ASP (Appletalk Session Protocol: a client/server protocol) , DNA
SCP (Digital Network Architecture Session Control Protocol: DECnet session layer protocol)

Transport (end-to-end communication between hosts)


-Ensures reliable & large size (large number of packets) data connections between hosts.
-Segmentation & reassembly
-Multiplexing upper-layer applications (TCP protocol or socket)
-Session establishment & tear-down of virtual circuits
-Hides details of any network dependent information from higher layers
-Reliable transport option
-Flow control
-ACK to sender
-Not ACK, then resend
-Manageable data flow maintained

"Window size" - how many data-segments to send before ACK.

Network (Routing)
-Sending packets from source network to destination network (finding path)
-Accomplished by packet switching via routers
-Network map created & used
-Interconnect independent networks
-Use of layer 3 addresses prevents layer 2 broadcasts

Data-link (Framing)
-Delivery of data to device
-Translates into bits for physical layer

Copyright © 2004 cisconotes 1


-Formats data into data frames

Frame format:
-Preamble (start indicator)
-Destination address
-Source address
-Ethernet II: type field
-802.3: length field (assumed IPX for type) length is between here and FCS
-802.2: adds LLC info
-DATA
-FCS

Data-link’s two sublayers:

LLC sublayer – mitigates MAC sub-layer & network layer


-SSAP Source Service Access Point
-DSAP Destination Service Access Point
-802.2 is 802.3 with the addition of the above two fields
-This layer also has timing, flow control, some even have connection vs. connectionless

MAC sublayer – framing: builds from bits at physical layer


-Checks CRC, then MAC address, if ok, passes up to next layer.
-Media access (part driver, part hardware): Contention (i.e. CSMA/CD), Token Passing, Polling
-First two bytes of Ethernet MAC address are unique per manufacturer. Can be changed in many NICs.

WAN Data-Link Protocols


-These are all layer 2 protocols, but their ability to bind to Layer 3 protocols varies.

HDLC
-Point-to-point or point-to-multipoint
-Most vendors implement HDLC differently
-Default for Cisco
-Cisco’s HDLC only compatible with Cisco

SDLC
-Originally for IBM mainframe
-Polling media access method between PRI & SEC

LAPB
-Use with X.25
-Defines frames & out-of-sequence, missing frames
-Frame: retransmission, exchange, acknowledgement

X.25
-point-to-point between DTE & DCE
-Supports SVCs & PVCs
-Precursor to frame-relay

SLIP
-Low speed serial for TCP/IP

PPP
-Adds login, password, error correction (on top of SLIP)
-Adds IPX, and Appletalk (on top of SLIP)

Copyright © 2004 cisconotes 2


Frame-Relay
-Upgrade from X.25 without LAPB
-Fastest wan protocol due to simple framing (no error correction)
-Uses SVCs, PVCs, DLCIs

Physical (last layer - transmitting bits via signaling and physical protocols across actual media)
EIA/TIA 232 or 449 V.24 V.35 X.21 G.703 EIA-530 HSSI – High-Speed Serial Interface
Interface between DTE and DCE is defined.

Ethernet & 802.3


-Ethernet_I (DIX) 1980
-Ethernet_II 1984
-Logical bus topology
-10Base2/Thinnet 185m RG58 coax 50 ohms
-10Base5/Thicknet 500m RG8 or 11 50 ohms
-10BaseT/UTP Cat 3 10Mbps, Cat 5 100Mbps, Cat 6 155Mbps, Cat 7 1Gbps

-some hubs are “stackable”—connect to hub backplane via wide cable to “upstream hub” to avoid collision
logic present in ports. This will sometimes turn off the downstream hub lights. The uplink port can also be
used to stack hubs to avoid collision logic. The uplink port is still crossed like all other hub ports, but
might have a button to switch it to uncrossed MDI/MDX button.

CSMA/CD
-Carrier Sense Multiple Access with Collision Detect
-Solution to problem when Ethernet frames are transmitted simultaneously from more than one NIC
causing the frames to be garbled.
-Routers and bridges prevent transmission of Ethernet frames throughout the whole network.
-Multiple Access refers to the capability of any NIC to being transmitting a frame at any time.

CSMA/CD Steps for a NIC


1 – Carrier Sense: see if the Ethernet digital signal is on the wire.
2 – If no other NIC is transmitting (Collision Detection), then the NIC will transmit its frame.
3 – That NIC continues to check to see if no other NIC is starting to transmit a frame.
4 – If that NIC hears a NIC starting to transmit a frame, then a long jam signal causes all nodes on the
segment to stop sending frames, then all nodes wait a little before trying to send frames again.
5 – After 15 consecutive collisions, nodes will time-out.

Ethernet Broadcasts
-Ethernet broadcasts are frames addressed to the following destination MAC address: FF:FF:FF:FF:FF:FF
-This means that the frame is addressed to all NICs.
-IP, and all other network layers, have their own form of broadcasts.

FDDI
-Dual token ring 100Mbps over fiber
-Very long distances
-Good in electronically hostile areas
-High speeds
-1 ring in use at a time
-2nd ring is backup (goes in opposite direction)
-Some stations attached to both rings, which is a DAS – Dual Attached Station; They are usually servers.
DAC - Dual Attached Connection is the term used in the same situation, but a server is not used. Most are
only SAS – Single Attached Station)
-Cisco routers can attach via dual-homing
-Logical and physical ring

Copyright © 2004 cisconotes 3


-(12) 4-bit symbols for addressing
-Permits several tokens

-High latency between Ethernet and FDDI conversion.


-No full-duplex
-Very expensive

-No tokens from NAUN , beacons (alert & locates failure)


-If not receive beacon, primary ring wraps to secondary ring at that point.

Token Ring
NAUN - Nearest Active Upstream Neighbor (this is the very last station that receives the token before the
station in question receives the token)
MSAU - Multi Station Access Unit (acts as a central physical hub for Token Ring)

-Star physical topology.


-Logical ring topology: Tokens are passed around in a circle from station to station.
-STP wiring.
-2 types: IBM & 802.5
-More expensive than Ethernet. More resilient under heavy loads.
-Each station receive signals from NAUN.

Token Ring Frame Header


-Many fields in the frame are the same as in Ethernet’s frames, i.e. Preamble, FCS.
-Two special fields, unique to Frame Relay follow:
Access Control Field
-3 Priority bits
-1 Token bit
-1 Monitor bit
-2 Reservation bits
-1 Priority bit

Frame Status Field


-A Address bit
-C Copied bit
-2 reserved bits
-A Address bit (for optional secondary destination)
-C Copied bit (for optional secondary destination)
-2 reserved bits

-Each station waits for token frame with empty T-bit, then appends data to frame and sets T-bit (if has data
to send): this prevents collisions entirely.
-Receiving station copies frame, then tags frame as copied. When frame reaches originating station, frame
is removed.
-Usually only one frame on ring at a given moment is max. But, can use early token release to transmit
new token immediately after transmit.
-Priority bits: only stations equal to or greater than in priority bits can claim the token.
-Only stations with higher priority can reserve token for next pass.
-Stations must reinstate previous priority after transmitting is complete.
-Frame status field includes A & C bits (Address bit, Copied bit). Both are cleared when sending (also
means destination not found yet). When sending station receives frame back, bits are read to ensure data
was received correctly. A, C bits respectively:
-1,1 – station found (data read by station) or frame copied to another ring (via a bridge)
-0,1 -- (2nd bit, the copy bit, is on) frame was copied by the destination station, but not
acknowledged

Copyright © 2004 cisconotes 4


-1,0 – The destination station was not able to copy the data from the frame.

Active monitor – a station that makes sure no more than 1 token, removes token from failed transmission
station

ATM
-Used for interactive media, real-time video, client/server databases, interconnection of existing networks.
-High-bandwidth, low delay
-Uses switching & multiplexing
-53-byte cells
-Works in both LANs & WANs
-Can allocate bandwidth on demand
-Not dependent on physical layer, but req. high-speed, high-bandwidth medium like fiber optics.

LAN Segmentation
-Bridge: 20-30% latency
-Router: 30-40% latency for ACK protocol 20-30% for sliding-window protocol
-Manageability – multiple routing protocols
-Increased functionality – flow, error, congestion control.
-Multiple active paths: using protocol, DSAPs, SSAPs, and path metrics to make informed routing
decisions

LAN Switching Types


Port switching – port is assigned to a segment
Frame switching – increases bandwidth, multiple transmissions in parallel (catalyst switches)
Cell switching – like frame switching, uses small fixed-length cells (Lightstream switches)

Switches
-Higher port density at lower cost than bridges.
-Layer 2 or layer 3 switches
-Numerous simultaneous conversations
-High-speed data exchanges
-Low-latency, high frame-forwarding rates
-Dedicated communication between devices
-Full-duplex communication
-Media rate adaptation 10/100 hosts on same network
-Works with 802.3 NICs & cabling

Switching Modes
Store and forward (a primary type) : copies entire packet, computes CRC. Discards if: CRC error, runt,
giant. Otherwise forwards (Catalyst 5000 series)
Cut-through (a primary type) : copies only destination address, then forwards.
Can be configured to operate below an error threshold on a per-port basis.
Fragment-free - waits for collision windows (64 bytes long). Errors almost always happen in 1st 64 bytes.
Better error checking than cut-through with almost no increase in latency

Multilayer Switching
-Switching based on layer 3 or even layer 4 address.

Layer 3 Switching
-Checks source and destination IP addresses, and switches accordingly

Copyright © 2004 cisconotes 5


-Less expensive & faster than routers because ASICs (Application Specific Integrated Circuits), which are
cheaper than router's processors.
-Routing protocol for layer 3 (i.e. RIP, OSPF) with connectivity to routers possible.
-Layer 3 fields are updated such as TTL (Time To Live)

Layer 3 Switching Options


-Filtering available based on policies.
-CRC at layer 3 is checked
-SNMP MIB updates
-QoS (Quality of Service)
-Traffic flow monitoring
-Most support IP only or IPX only. Few support both.
-Limited media support.
-May replace collapsed backbone routers in high-speed LANs
-Supports IP & IPX.
-Supports fewer media types than routers.

Layer 4 Switching
-Checks layer 4 ports, i.e. TCP port # or UDP port # and switches accordingly)
-Less expensive and faster than most routers because ASICs are used.
-all layer 3 switch features
-QoS (Quality of Service) on per user or per application basis.
-memory requirements large since every port# is stored!

Cisco MLS (Multi-Layer Switching)


-Rewrite layer 2 & 3 headers at wire speed (via ASICs). Only the very first packet from a given IP makes
it to the RP, all subsequent packets from the same IP is cached, which results in layer 2 & 3 headers being
rewritten at wire speed.
-MLS also known as "route-once, switch many."

Spanning Tree Protocol (STP - 802.1d)


-Prevents routing loops, allows redundant network path only in case of backup (due to failed link).
-STP detects, and then blocks redundant paths (aka loops).
-STP does this by calculating a good spanning-tree topology.
-STA implemented by STP prevents loops by calculating a stable spanning-tree network topology; done via
BPDU (Bridge Protocol Data Units)

VLAN (Virtual LAN)


-Logical grouping of users & resources (i.e. location, function, dept.)
-Connected to defined ports on the switch
-Looks and acts like a subnet
-Simplify moves, adds, changes
-Reduce administrative costs
-Better control of broadcasts
-Tighten network security
-Microsegment with scalability

Frame Tagging
-Unique user-defined ID is placed in header of each frame (used only for identifying source VLAN).
(ONLY exists while in switch fabric (trunk lines)) Exception: ISL (Inter-Switch Link).
-ID is examined by each switch prior to any broadcasts or transmissions.
( 3000, 5000 series switches)
-Developed for multi-VLAN inter-switch communication
-Little processing or administrative overhead

Copyright © 2004 cisconotes 6


Half-Duplex Ethernet
-Transmit circuits active in sending NIC, it’s transmit wire, and receive circuits active in other NIC.
-Loopback and collision detection circuits active on both NICs; Listens for activity on 2nd wire: if present
then collision is detected: both NICs will pause.

Full-Duplex Ethernet
-Requires a switch port on “hub”.
-Can be used in 10BT, 100BT, 100BFL.
-Point-to-point -> collision free: doesn’t share bandwidth.
-Uses dedicated transmit wire for each NIC port: no chance of collision.
-Requires full-duplex NIC card & S/W drivers must support 2 simultaneous data paths.
-Requires Loopback & collision detection disabled on hub (makes it a switch port).
-Up to 200Mbps max (if both parties transmit 100Mbps at same time).

Crossover Ethernet Cabling


-Straight-through cables (normal cables) are wired from pin1 to pin1on the other end, the same goes for all
the other pins. They are used for connecting hosts to hubs (or switches).
-Crossover cable has transmit pins connected to receive pins, and vice versa. This allows connecting two
hosts directly together without a hub. They are used for:
-Uplinks between switches
-A hub to another hub
-Hubs to switches
-Easily distinguish Ethernet cable between crossover cable vs. straight-through wired cable by placing the 2
connectors side by side, facing the same way. If the color bands are in the same order from left two right,
for both cables, then that cable is a straight-through wired cable.

MDI/MDX Buttons
-Medium Dependent Interface
-MDI-X ports is normally what all hub (or switch) ports are permanently set to (no MDI/MDX button to
change it's type). This is because hubs (or switches) connect the TX to RX and vice versa.
-MDI ports is normally what all PC Ethernet ports (and router Ethernet ports) are permanently set to (no
MDI/MDX button to change it's type).
-MDI-X port connected to MDI port: use a straight-through connecting cable.
-MDI-X port connected to another MDI-X port, or MDI port connected to another MDI port: use a
crossover connecting cable.
-MDI/MDX autosense is available on some Cisco products

Fast Ethernet
100BaseT Fast Ethernet – 802.3u
100BaseFX Ethernet over fiber 100Mbps using 802.3 specs. Two strand 50/125 or 62.5/125 –micron
multimode fiber-optic cable.
100BaseT4 4-pair 100Mbps using 802.3 specs over CAT 3,4,5. RJ45 connector.
100BaseTX Fast Ethernet over CAT 5. To 802.3 specs. Alternate: 2 pair, 100 ohm STP or type 1 STP.
100BaseX means either 100BaseTX or 100BaseFX. Std approved for compatible: CSMA/CD& ANSI
X3T9.5
100VG AnyLan not compatible with 802.3 standards, Cisco does not support.

Advantages of Fast Ethernet


-Same signaling as 10BT: gradual migration possible.
-10X performance of 10BT
-Existing cabling and network equipment can be used
-10 and 100 can be used together

Copyright © 2004 cisconotes 7


-Tried/tested CSMA/CD technology
-Easy migration

100BT Specifications
100BT uses same time slots as 10BT
-Round-trip must not exceed 512 bit-times for ANY Ethernet implementation. Longer times means too
many collisions than allowed by normal Ethernet operation.
for fast Ethernet, this means shorter distance
-Only 2 class II repeaters max. Not 5-4-3 rule!

100 BT Repeaters:
-100BaseFX repeaters reduce max distances due to slow conversion.

CLASS I – translational repeater support: 100 base X & 100 base T4


Has 140 bit time delay

CLASS II – transparent repeater: Either 100 base X OR 100 base T4


Has 92 bit time delay

FastHub 300 – repeater compatible with IEEE 802.3u. 10X performace of 10BT hub. Slightly better than
previous two at max distances: A single repeater always has 200m max dist (same 10BT cable restriction of
100m) . Best combination : 2 FastHubs: 223m (100m cable length still in force for UTP). TX/FX
combination on hub(s) 216-308m.

Max, dist between end nodes


Port Type
100BaseTX Cat 5 RJ45 100m
100BaseFX 50/125 or 62.5/125 SC/ST/MIC Half Duplex 412m Full duplex:~1Km(attenuation only)

DoD Reference Model


-Condensed version of OSI model

DoD OSI

Process/ Application Node-to-Node application communication, controls user


interface specifications
Application Presentatiuon

Session
––––––––––––––––––––––––––
Host to Host Transport Protocols for setting the level of transmission service for
applications. Create reliable end-to-end communication and
ensuring error-free data and packet sequencing.
––––––––––––––––––––––––––
Internet Network Logical transmission of packets over entire net. Adressiong of
hostst, hadles routing of packets across multiple networks, controls comm. Flow between 2 hosts.
––––––––––––––––––––––––––
Network Access Data Link Monitors data exchange between host and network. H/W
addressing. Defines protocols for physical transmission ofdata
Physical

-DoD model requires how done to be different than OSI model.


-Implies DoD model has different suite of protocols.
-Most applications written with TCP/IP protocols are client/server applications.

Copyright © 2004 cisconotes 8


TCP/IP Protocol Suite
-Protocols are found in both the DoD and OSI models.

Telnet
-Terminal emulation
-Telnet client access resources of telnet server
-Make telnet client appear as though it was a terminal attached to local network
-It’s a software image (virutal terminal) good for running apps or snooping around

FTP
-Uses telnet for transparent log in
-As a program: file/directory listing / manipulation / viewing

TFTP
-Transfers files with smaller blocks.
-No authentication.
-Need path/filename, web or IP address.

NFS
-File sharing for Unix
-Allows 2 different types of file systems to interoperate
-i.e. NFSserver on NT can provide Unix & NT files
-NFS can: execute remote files as programs
-Import & export material
-Manipulate apps. Remotely

SMTP
-Delivers email by queueing emails (usually to disk) on server from a host

LPD – Line Printer Daemon – queue for printer sharing


LPR – for printing (uses lpd unix command)

X-Window
-Defines protocol for writing GUI client server apps.
-“Client” program on one host
-“Window server” program on another

SNMP
-Collects & manipulates valuable network information.
-Polls devices on netwrok from network management station
-When network is healthy, SNMP creates baseline
-Allerts managers to sudden changes
-Agents send alerts, “trap” to management station
-Agent’s threshold can be set usu. By examining baseline
-Software that uses SNMP:
-CiscoWorks (LAN/WAN monitoring of routers & switches))
-CWSI (CiscoWorks for Switched Networks) Monitoring for switched internetworks (monitoring
& managing switches & VLANs)
-NetCool

Host-to-Host
Shields upper layers (applications) from complexity of network
Creates virtual circuit (connection oriented)
Agree on size of window.

Copyright © 2004 cisconotes 9


TCP
-Ensures reliable connections between hosts.
-Test for errors, resends data if necessary, reorts the occurence of an error to upper layers if it can’t
manage to solve the problem itself.
-Breaks large blocks from applications into segments
-Waits for segment ACK’s after sending segments
-TCP contacts destination’s TCP to establish a connection known as a virtual circuit (connection- oriented)
-Agree on window size during this establishment phase.
-Full duplex, connection-oriented, reliable, accurate protocol.
-Very complicated, adds much to network overhead.
-Today’s networks very reliable, so TCP often unecessary

TCP header format


Source Port 16 bits
Destination Port 16 bits
Sequence number 32 bits
Acknowledgment number 32 bits
HLEN 4 bits
Reserved 6 bits
Code Bits 6 bits
Window 16 bits
Checksum 16 bits
Urgent pointer 16 bits
Option 0 or 32 bits
DATA variable

Source Port – port or socket of the application that sent the data
Destination Port – port or socket of application on the destination host
Sequence Number – used to put the data back in the correct order or retransmit missing or damaged data
Acknowledgment number – defines which TCP octet is expected next
HLEN – header length defines the number of 32-bit words in the header
Reserved – always set to zero.
Code bits – sets up and terminates a session
Window – the window size that the sender is willing to accept
Checksum – CRC, TCP doesn’t trust the lower layers.
Urgent pointer – indicates the end of urgent data.
Option – sets the maximum TCP segment size
Data – data handed down to TCP protocol.

TCP port numbers (applies to UDP port numbers also)


-Used to keep track of different conversations crossing the network simultaneously
-Originating source port numbers are dynamically assigned by the source host, usu. > 1023

well-known port numbers (RFC 1700)


0 – 255 public applications
255 – 1023 for companies to use in their applications
> 1023 set up sessions with other hosts (for originators), & as source, destination TCP addresses
these are dynamically assigned when creating TCP virtual circuits.

UDP
-Scaled, down, economy model version of TCP (thin protocol) Æ lower overhead
-Much fewer options than TCP
-Unreliable
-Used with SNMP to avoid using up network bandwidth with TCP protocol overhead. SNMP are small
messages, no need for TCP to be used.

Copyright © 2004 cisconotes 10


-Some Process/App protocols handle reliability, such as NFS.
-Use of UDP decided on by application developer.
-UDP receives upper layer blocks, instead of streams of data like TCP, and breaks them into segments.
-Does NOT sequence segments, order unimportant
-No checkup, no follow-up.
-Does not create virtual circuit
-Does not contact destination before delivering information (connectionless)

UDP Segment format


Source Port 16 bits
Destination Port 16 bits
Length 16 bits
Checksum 16 bits
Data variable

TCP UDP
Reliable Unreliable
Virtual circuit Connectionless
Sequenced Unsequenced
Acknowledgments Low overhead

Internet Layer Protocols


-Routing & providing a single network interface to upper layers
-IP is the Internet layer, other protocols support it.
-Sees all interconnections between networks, has logical software address (IP address)-
-Checks IP address, consults routing table for best path.
-Allows software developers to ignore the network types (Ethernet, Token Ring) Æ Single network -
interface for developers.

IP (Internet Protocol)
-Packets are used on this layer.
-Which network is it on, what’s it’s ID on that network -> logical address (IP), hardware address
-IP address used for routing, each machine reads IP source, IP destination, routes based on these two.
-Fragments segments into packets (aka datagrams), and packets back to segments,

IP header
-Version: IP version number
-HLEN: header length in 32-bit words
-TOS: Type of Service tells how the datagram should be handled
-Total Length: the length of the packet including header and data.
-Identification, Flags, Frag Offset: These provide fragmentation and reassembly if the packet is too large to
put in a frame.
-TTL: Time to Live.
-Type or Protocol: port of upper-layer protocol ( TCP – port 6 , UDP – port 17)
-Header checksum: Cyclic Redundancy Check or Frame Check Sequence (FCS)
-Source IP Address: 32-bit IP address of sending station.
-Destination IP address: the 32-bit IP address of the station this packet is destined for.
-IP Option: used for network testing, debugging, security (i.e. PPTP encryption used in VPNs), and more.

RARP (Reverse Address Resolution Protocol)


-Determines an IP address for a host that does not have one (i.e. diskless workstation).
-RARP server responds with the IP address for the host when the host sends out it’s MAC address.

BootP (Bootstrap Protocol)


-Diskless workstation issues BootP broadcast on the network.

Copyright © 2004 cisconotes 11


-BootP server responds by looking up MAC address in BootP file.
-BootP server tells the machine it’s IP address & file it can boot from (usually via TFTP)

ICMP (Internet Control Message Protocol)


-Management and message service for IP.
-Messages carried as IP datagrams
-Has annexed ability: Router Discover Messages (finds gateways for hosts). This is done via router
advertisements and router solicitation (request immediate advertisements).
-"Destination unreachable": router can’t send IP datagram any further (this is sent back to sender)
-"Buffer full": router’s memory buffer for receiving datagrams is full.
-Hops: number of routers an IP datagram it may go through. The last router to receive this count of hops
sends back exceeded TTL to sender.
-Ping: Packet Internet Groper uses ICMP echo messages to check physical connectivity.

Network Access Layer Protocols


-Receiving an IP datagram and framing it into a stream of bits for physical transmission.
(i.e. CSMA/CD).
-Specifying the MAC address. Places MAC address into frame
-Ensuring bit stream making up the frame has been accurately received (CRC check)
-Specifying access method to physical network (contention-based, token-passing, polling)
-Specifying the physical media, connectors, electrical signaling, and timing rules.

Frames
-Ethernet_II : source, destination, type
-802.3: source, destination, length (IPX always assumed for type, but can be bound to IP instead). Novell
originally used 802.3.
-802.2 is an 802.3 frame with DSAP, SSAP fields (LLC sublayer).
-SSAP Source Service Access Point is used to define which network layer the source of the frame is using.
-DSAP Destination Service Access Point is used to define which network layer the frame's destination is
intended to go to.
-MAC sublayer has error detection, LLC sublayer has error correction (more efficient)
-Examples of implementation of this layer on LANs: Ethernet, Fast Ethernet, Token Ring.
-Examples on WANs: Point-to-Point Protocol (PPP), X.25, Frame Relay

ARP and the ARP Cache


-ARP is used by IP whenever the destination MAC address is unknown.
-ARP uses an Ethernet broadcast to determine destination MAC address (belonging either to a host on the
same network or the default gateway’s (usually a router) MAC address) from a given destination IP
address.
-Once an IP address is resolved into a MAC address, it’s recorded on the host’s ARP cache.

ARP steps
1. Destination IP checked to see if it’s local (via subnet mask). If local, the host system checks it’s ARP
cache for the MAC address

2. If not found in ARP cache, host sends Ethernet broadcast on subnet asking for MAC address with
given IP address.

3. A reply (containing the needed MAC address) from the host with that IP address is sent directly to the
hardware address of requesting host.

4. Requesting host appends this IP address to ARP cache. Communication can now take place since the
destination host’s MAC address is known.

-To see contents of ARP cache on a Windows machine, use arp –a from a DOS prompt.

Copyright © 2004 cisconotes 12


-Use show arp on a Cisco router or Cisco switch.

Note: ARP entries on Windows last minutes. ARP entries in Cisco last for hours.

IP Addressing
-Hierarchical address (Net:Node), not "flat address". 32-bit (4 octets).

Class
A Net.Node.Node.Node 0 127
B Net.Net.Node.Node 10 128-191
C Net.Node.Node.Node 110 192-223

D multicast packets 224-239


E reserved

Private Addresses
-These are not routable on the Internet by default. They are used in private networks, typically behind a
Proxy server. They are also frequently NAT'ed (discussed later).
Class A: 10.0.0.0
Class B: 172.16-31.0.0
Class C: 192.168.0.0

Global IP address - IP addresses that are routable on the Internet

network address all 0 this network or segment


network address all 1 all networks

network 127 Loopback tests. Designates the local node, and allows that node to send a test
packet to itself without generating network traffic.

node address all 0 this node


node address all 1 all nodes on specified network.

entire ip address set to all 0s For Cisco routers, means default route.

entire ip address set to all 1s Broadcast to all nodes on the current network.

Subnetting a Nework
-Uses bits from host portion of an IP address to create a subnet.
-Subnetting is act of creating little subnetworks from a single, large parent network.
-An organization with a single network address can create a subnet for each physical network (most have
multiple physical networks, so this solves that problem)

Benefits
-Reduced network traffic
-Optimized network performance
-Simplified management (easier to find problems in a smaller group of networks)
-Facilitated spanning of large geographical distances.

Problems if there was no subnetting


-Single network address can be used to refer to multiple physical networks, if a network address for each
physical network was granted, not enough addresses would result.
-Router’s routing tables would be impossibly large (each machine on every network in a company would
have to be known to each and every router in the company)

Copyright © 2004 cisconotes 13


-Routing information exchange would cause too much network traffic

Planning Subnetting
-Determine current requirements and future conditions

Determine required network IDs


-One for each subnet
-One for each wide area network connection

Determine required host IDs per subnet.


-One for each TCP/IP host
-One for each router interface

Based on the above requirement, create


-One subnet mask for your entire network
-A unique subnet ID for each physical segment (router segmented)
-A range of host IDs for each subnet

Subnet Masks
-Each machine on the network must know which part of the host address will be used as the subnet address.
-A subnet mask is assigned to each machine to do this.

-1’s refer to network or subnet address. 0’s refer to host part of the address.
-Default subnet masks are for networks without subnets, i.e. Class B network defaults to a 255.555.0.0
subnet mask. Also, when subnetting, this is the subnet mask’s minimum size.
-Invalid subnets (those that don’t have the minimum size) typically aren’t even allowed to be punched into
IP software.

-Normally, total number of subnet bit positions, 1’s, results in total number of possible subnets: 2^bits -2
-The -2 is there because all 1’s in last octet and all 0’s in last octet are reserved. There is a special Cisco
command to allow all 0’s ip subnet-zero... This will allow both the highest and lowest possible subnet
combinations as valid subnet numbers. Make sure all IP software recognizes this convention before using.

-Masked positions are 1’s, and unmasked are 0’s.


-Total possible hosts are calculated using same formula as above except uses unmasked bits.
-If the subnet mask is too large, resulting in fewer possible hosts, it is possible to add a secondary IP
address to a router interface (but this reduces the possible subnets).

-Calculate valid subnets: 256 – subnet mask = first subnet. Keep adding first subnet to itself until subnet
mask reached (discarded).

-Incorrect mask on a workstation can cause the router to assume the workstation is on a different subnet...
-The result is the router won’t forward packets to it. Also, the workstation will send to the default gateway
when it shouldn’t.

-Calculate valid hosts: for each subnet, it’s all the numbers above the subnet number, minus the all hosts
broadcast.

-Very large subnet masks such as 255.255.255.252 (2 host) are common in networks with many WAN
links. They usually have a direct connection between each site, so only need two host addresses, one for
each router port. Only EIGRP, OSPF support VLSM, one subnet for the LAN, and a different one for the
WAN. Only EIGRP, OSPF transmit subnet masks in their routing tables.

Cisco Routers
-Configure by console port, aux port.

Copyright © 2004 cisconotes 14


-Those are both asynchronous serial ports. A console port connection is connected to by laptop or
PC running terminal emulation s/w or a console. The aux port can have a modem attached to it.
-Configure by any other interface. So long as the interface is connected to a network, IP protocol
configured correctly, and port is active.

-The console port is used initially, then can telnet into an interface like e0.

-Router can be set up via external sources:


-Virtual terminals
-Network management stations

Router Components
ROM: stores the bootstrap program, operating system, and POST. They are socketed for upgrade or
replacement

Flash: EEPROM that holds the operating system image and microcode. Allows upgrades without
replacing the ROM chip.

RAM: provides caching and packet buffering, routing tables, etc. Holds the operating system when the
router is turned on.

NVRAM: stores the router’s startup configuration file.

Interfaces: either on the motherboard or as separate modules. ISDN interfaces can be installed in the
router separately.

Startup Sequence
POST (Power On Self Test) - Checks the CPU, memory, and all interface circuitry.

1) Bootstrap program is stored in and executed from ROM on the CPU card, it searches for a valid
CISCO IOS image (from flash, TFTP server, or ROM)

2) Operating system source is determined from boot field setting in the router’s configuration register.
If flash or network load, boot commands in the configuration file specify exact location of the image.
FLASH is default.

3) Operating system image is loaded into low-addressed memory. Determines h/w, s/w components and
lists this on screen.

4) Configuration file in NVRAM is loaded into RAM, then executed line by line.

5) System configuration dialog (aka setup dialog) started if no valid configuration file exists in NVRAM.
Can be configured by console or TFTP at this dialog.

Exec Command Interpreter


Two levels (modes) of access: user and privileged.

User mode: checking router’s status, viewing basic system information, connecting to remote devices,
temporary changes to terminal settings.

Privileged mode: used to change the configuration of the router. All commands in user mode, plus those
used to set operating system parameters, detailed information on router’s status, test and run debug
operations, and access global configuration modes.

Copyright © 2004 cisconotes 15


Manual Configuration
In privileged mode (# prompt):

config t – configure via console port or telnet.


config mem – configure via copying startup-config to running-config
config net – configure via a network TFTP server.

Logging into router

After first logging into router, user Exec mode prompt: >
? for a list of commands any time.

-- more -- .. press return to view line by line. Press space bar to see the next screen.

logout to exit user mode.

enable to enter privileged mode. (asks for password)

disable to return to user mode.

exit & quit take you out of both the privileged and user modes and log you out of the router.

Help on any command by typing ? after the command, i.e. clock ?

cl? Lists all command starting with cl

clock ? lists next keyword or argument needed.

terminal editing helps type in repetitive commands (commands listed below)

terminal no editing (this is the only command without no before the rest of the command)

-Terminal editing also scrolls the typed input to the left 10 characters when reaching column 78. When it
scrolls, it puts a $ on the very left of the line (after the prompt)_to show there is some missing on the left
side.

Terminal Editing Commands


CTRL+A Move to the beginning of the command line
CTRL+E Move to the end of the command line
CTRL+F (or rt arrow) move forward one character
CTRL+B (or lt arrow) move left one character
CTRL+P (or up arrow) repeat previous command entry
CTRL+N (or dn arrow) next (more recent) command recall

ESC+B move backward one word


ESC+F move forward one word

show history show command buffer


terminal history size set command buffer size

TAB completes entry typing for you.

Router Commands
-Two configurations: startup and running.

Copyright © 2004 cisconotes 16


-Startup is held in NVRAM, and accessed when the router is started and placed into DRAM.

show startup-config configuration file loaded when router boots.

-Running config is the same as the startup config except it includes all config changes since the router was
powered up.

copy running-config startup-config

show running-config shows the configuration that’s currently loaded into RAM and running.

erase startup-config erases the configuration in NVRAM, and puts it right back into initial configuration
dialog.

reload reloads the startup-config into memory.

setup this starts the initial configuration dialog.

IOS v10.3 and Earlier Commands


show config same as show startup-config
write term same as show running-config
write erase same as erase startup-config
write mem same as copy running-config startup-config

These commands won’t be available in future IOS versions.

Passwords
enable secret password for the enable prompt that is encrypted. Has precedence over enable password
when it exists. Configure the password in global configuration mode or setup mode.

enable password is used when there is no enable secret and when using older software and boot images.

-Virtual terminal password is required for telnet sessions into the router.

In global config mode:


(config)# line vty 0 4 – specifies the number of telnet sessions allowed. Different password for each: line
vty [port number]
(config-line)# login
(config-line)# password mypassword
(config-line)# ^Z

Auxiliary password
(config)# line aux 0
(config-line)# login
(config-line)# password mypassword
(config-line)# ^Z

Console password
(config)# line con 0
(config-line)# login
(config-line)# password mypassword
(config-line)# ^Z

banner motd # Message of the day banner displayed whenever anyone attempts to login.

hostname [name] this changes the name of the router, it’s displayed at each prompt.

Copyright © 2004 cisconotes 17


-Interfaces (aka ports) are specified like this: s0, s1, e0, to0, f0. For serial #0, serial #1, Ethernet #0,
Token Ring#0, FDDI #0.

-Switch to interfaces configuration mode via: int [interfacetype_andnumber]

shutdown turn the administrative state of the interface down.

description add a description or text line to an interface

-If modular (i.e. catalyst 5000 switch), syntax is: type slot/port. i.e. e 4/2 meaning card in slot 4, port 2.

-Versatile Interface Processor (VIP) cards: one or two slots for each port adapter, each port adapter has
interfaces. They are found in 7000 and 7500 series routers.

-Type slot/port adapter/port (interface)


-i.e. e 2/0/1 Ethernet adapter in third slot, first port adapter, second Ethernet interface.

Erase startup-config followed by reload results in system configuration dialog screen.

System Configuration Dialog

Would you like to enter initial configuration dialog? [yes]


Would you like to see the current interface summary? [yes]

Enter host name [Router]:


The enable secret is a one-way cryptographic.....
Enter enable secret:
The enable password is used when ...
Enter enable password:
Enter virtual terminal password:
Configure SNMP Network management? [yes]:
Configure IP? [yes]
Configure IGRP routing? [yes]
Configure RIP routing? [no]

Configuring interface parameters:


Configure interface ethernet0:
Is this interface in use? [yes]
Configure IP on this interface? [yes]
IP address for this interface:
Number of bits in subnet field [0]:
Class B network is 172.16.0.0, 8 subnet bits; mask is 255.255.255.0

Configure interface Serial0: (same as above, except:)


Configure IP unnumbered on this interface? [no]: (after it asks to configure IP)

Configure interface... then shows startup-config file


Use this configuration? [yes/no]

DTE/DCE Cable
-Normally, Cisco routers default to DTE devices.

-CSU/DSU normally handles the clocking for synchronous communications. They normally plug into
Cisco serial ports, but can be integrated into the router.

Copyright © 2004 cisconotes 18


-Without CSU/DSU, connecting routers directly using DTE/DCE cable, you must specify the clock rate
command, this simulates a DCE device.

Bandwidth is used by routing protocols (IGRP) to make routing decisions.

-Cisco serial port has 60 pins, some pins are looped to distinguish between DTE or DCE cable.
show controllers serial [port number] shows whether DTE or DCE cable is attached.

-If three routers with serial connections between them are connected in series, the middle router has two
DCE serial interfaces. The other two routers have DTE serial interfaces.

-DCE is a modem or interface card used to connect end-user equipment (DTE) to a greater communication
network composed of DCEs.

RouterA#sh controllers s 0
HD unit 0, idb = 0x0011d1, driver structure at 0x101de0
buffer size 1524 HD unit 1, V.35 DTE cable
cbp = 0xE1, eda = 0X3140, cda = 0x3000

RouterA#sh controllers s 1
HD unit 1, idb = 0x003de1, driver structure at 0x11ee0
buffer size 1524 HD unit 1, V.35 DCE cable
cbp = 0xE2, eda = 0X2840, cda = 0x2700

Routing Basics
IP routing – the process of sending data from a host on one network to a remote host on another network
through a router or routers.

Routing table contains the IP addresses of router interfaces that connect to the other networks.
Routing table is consulted for a path to the network that is indicated by the packet’s destination address.
If a path isn’t found, the packet is sent to the router’s default route address (if configured).

(Know how a hop across a router is accomplished by changing MAC address at the router)

Initially, a router only knows how to reach networks or subnets directly connected.
Then a router learns new networks by:
Static routing
Default routing
Dynamic routing

Static routers – require routing tables to be built and updated manually. Static routers do not
automatically share information when a change occurs.

Benefits of Static Routes


-Bandwidth conservation; no broadcasts for routing table updates are used.
-Security; routers only know about networks that you want them to know about – i.e. prevent things like
browsing the internet.

Tips: -Limit static routes to remote networks to only 1 router away.


-Use static routes when a network is reachable by only one path (stub network); avoids dynamic
updates.

In global configuration mode:

ip route network mask address | interface [distance]

Copyright © 2004 cisconotes 19


-Network: destination network or subnet
-Mask: subnet mask
-Address: IP address of next hop router
-Interface: name of interface to use to get to destination network
-Distance: the administrative distance- source’s trustworthiness – 0 to 255. 255 is least trustworthy.

show ip route
C 172.16.50.0 is directly connected, Ethernet0
S – static I- IGRP, R-RIP, M – mobile, B – BGP, D-EIGRP, EX – EIGRP external, O – OSPF...
-This shows one directly connected route.

show ip route static


S 172.16.50.0 is static, Ethernet0
S – static I- IGRP, R-RIP, M – mobile, B – BGP, D-EIGRP, EX – EIGRP external, O – OSPF...
-This shows one static route.

Default route
-aka router or gateway of last resort)
-If the router doesn’t know the path to a certain network, it can use it just like a host uses a default gateway.

ip route 0.0.0.0 0.0.0.0 172.16.20.2

-Set address, and subnet mask to all 0’s.


-Even if routing tables aren’t set correctly, this would help, but problems may still be encountered!
-Best to have a combination of static and default routes.

ip classless must be used when using a default route for it to work.

-If a Cisco router doesn’t have the subnet’s number and there’s no network default route, packets are tossed
even though the packet falls within the subnet-addressing scheme.
-To get the Cisco IOS to forward packets to obscure subnets, use ip classless.
-In other words, it tells Cisco IOS to forward the packet to subnets unknown to its routing table. It will use
a directly connected network or default route for the packet.

Dynamic Routing
-Ideal for large internetworks; maintaining static routes would be overwhelming.
-Minimal configuration required for dynamic routing.

Routed vs. Routing Protocols


-Routed protocols are Network layer protocols such as IP or IPX. They are used to direct user traffic via
the header information. Routed protocols specify the type of fields within the packet. Packets with routed
protocols can send data from sending host to receiving host.

-Routing protocols maintain the routing tables used between routers. Routing protocols don’t send end-
user data from network to network, they only pass routing information.

-Routers support multiple independent routing protocols, and can do this for all routed protocols
simultaneously.

-Many networks over the same media are possible this way (IP and IPX on same physical network).

-Most network communication occurs within small groups, routing systems follow this by creating logical
groups: domains, autonomous systems, or areas.

Copyright © 2004 cisconotes 20


-A routing domain or autonomous system (AS) is a portion of an internetwork under common
administrative authority. An AS consists of routers that share information using the same routing protocol.
-Some routing protocols allow subdivision of an AS into areas. Typically there are a few areas and one
area designated as backbone. Routing within a domain = intradomain routing. Routers that connect the
areas to the backbone are called: external, inter-domain, and inter-AS routers.

-If the internetwork is connected via Internet, the NIC will assign you a unique 16-bit AS number.

Interior Routing Protocols


-Implemented on internet layer
examples:
RIP – distance-vector routing protocol
IGRP – Cisco’s proprietary distance-vector routing protocol.
OSPF – a link-state protocol
EIGRP – Cisco’s balanced distance-vector routing protocol

Two basic categories of routing protocols: distance-vector and link state.

Distance-Vector Routing Protocols


Distance vector - knows direction and distance to any network connection. It listens to secondhand
information for updates.

-Convergence: time for all routers to update their tables when a change occurs.
-update every 30 (RIP) to 90 seconds. All routers pass entire routing tables to all other known routers.
-when routers power up, they get to know the neighbors: learn metrics (hops) to all other routers out int.
-Discover (via these updates) best paths to destination networks based on number of hops the routers are
from each neighbor.
-Whenever network topology changes, routing table updates occur by each router sending out its entire
routing table in the form of a broadcast to all other routers.
-Hop-counts used (only) when making routing decisions. Sometimes have to lie to router about hops
(change hop count) when a high speed line is available.
-Routing loops occur. Happens because every router is not updated at close to the same time.
-Routing loops also known as counting to infinity (exceeding TTL). This is the result of wrong information
being communicated and propagated.
-Distance vector permits a maximum hop count of 15 to keep packets from staying in a loop forever.
-Split horizon: reduces incorrect routing information and routing overhead in distance-vector network by:
not allowing information to be sent back in the direction from which it was received.
-Route Poisoning: When a directly connected network goes down, that router changes the table entry for
that network to show 16 or unreachable aka infinite. This stops incorrect updates about that network for
this router. This router will keep it in this state until the network is back up, at that point it will trigger an
update to be sent out (don’t have to wait 30 seconds).
-Hold Downs: Used with Route Poisoning to prevent regular update messages from reinstating a down
route. They use triggered updates, which reset the hold-down timer.

Link state
-Maintains a more complex table of topology information (complete understanding and view of all the links
of distant routers)
-Uses LSP link-state packets to inform other routers of distant links.
-Uses topological databases, the shortest path first SPF algorithm, and a routing table.
-Path choice based not just on hop count, but also: available bandwidth, and congestion.

Link State Protocol algorithm


-First, routers exchange LSPs with one another: includes only directly connected links.
-Second, routers compile all of the LSPs from the internetwork and build a topological database.

Copyright © 2004 cisconotes 21


-Third, SPF computes how each network can be reached: both shortest and most efficient paths to each
link-state network.
-Fourth, each router creates a tree structure with itself being the root.
-Finally, the results go into a routing table (without best paths).

-Handles convergence differently than distance-vector. When topology changes, routers that are aware of
the change first send information to all other routers with same link-state algorithm or to a special
consultative router (Designated router, DR).
-Whenever router receives LSP packet, router recalculates the best paths and updates the routing tables
accordingly.
-Consider issues of processing power, memory usage, and bandwidth requirements.
-Most bandwidth consumed when routers come on initially; regular updates sent every two hours on avg.
(configurable). Otherwise, broadcasts are sent only when a change occurs.
-LSP time stamps, sequence numbers, and aging schemes used to avoid spread of inaccurate LSP info.

Balanced Hybrid
Uses the best of both distance-vector and link-state algorithms.
-Uses distance vectors with more accurate metric counts.
-Uses link-state triggers
-More efficient link-state protocol (not so much processor, memory, and bandwidth required).

Ex: IS-IS, EIGRP.

RIP (Routing Information Protocol)


-RIP & IGRP summarizes routing information by major network numbers (classfull routing). Classless /
prefix routing protocols allow contiguous blocks of hosts, subnets, or networks to be represented by a
single route.

RIP tables have the following minimum information


-IP destination address
-Metric 1-15: cost in hops.
-IP address of next router
-Mark signaling recent changes to a route
-Timers – for performance
-Flags – whether information about the route has changed.
-Hold downs (prevent reinstatement of non-functional routes_
-Split horizon (prevent routing loops)
-Poison reverse update. (prevents larger routing loops)

-RIP sends routing updates and when network topology changes.

-RIP has slow convergence.


-RIP in a large network can lead to counting to infinity and routing loops.

-Link State, aka shortest path first: knows entire network better than distance-vector. Never listens to
secondhand information. More accurate and informed routing decisions.

-RIP good for small to moderate sized homogenous internetworks (due to small hop count and single
metric)

Command field 1 byte request or response (reply to unsolicited routing update)


Version number 1 byte version of RIP
Zero 2 bytes
Address family 2 bytes type of addressing scheme used by destination address (2 with IP)
identifier
Zero 2 bytes

Copyright © 2004 cisconotes 22


Address 4 bytes
Zero 4 bytes
Zero 4 bytes
Metric field 4 bytes hop count to destination

25 destination per RIP packet max.

RIP Timers
Routing Update Timer - interval between sending complete routing table to all neighbors. (usu. 30 sec.)
Route Invalid Timer - time before router considers a route invalid. (hasn’t heard any updates about the
route. In that time). Router will then send update to all neighbors informing of invalid route.
Route Flush Timer - time between route becoming invalid and removal from the routing table (usu. 240
sec.). Router informs all neighbors of routes impending flushing. Route invalid < route flush

Interior Gateway Routing Protocol (IGRP)


-Distance vector interior routing protocol, proprietary to Cisco.
-IGRP is answer to RIP’s shortcomings for larger autonomous systems.
-IGRP – 255 hop count max. RIP – 15 hop count max.

IGRP Routing Tables


-Metrics: 1-255 set by an administrator to influence route decision.
-Delay: The speed of the media in 10ms units. 10Mbps Ethernet: 100 (1 sec).
-Bandwidth: 1200bps to 10Gbps
-Reliability: fractions of 255 (255 is optimal)
-Load: fractions of 255 – saturation of link (0 is no load)

-All above can be used for routing decisions, and administrators can adjust IGRP settings.

-For stability, IGRP uses the following:


-Multipath routing: Dual lines of equal bandwidth running a single stream of traffic (taking turns bearing
the load). Results in performance and stability

Hold-downs - Prevents regular update messages from reinstating a downed link. If a link goes down
during convergence, it will prevent routers that haven’t heard of the news of the downed link from
misinforming other routers that they have a path to the route. Hold downs prevent any changes that might
affect stated routes for a specific period of time (slightly greater than time necessary to update the entire
network).

Split Horizons - information is never sent back to the router from which it came to prevent routing loops.
Hold downs prevents routing loops, IGRP uses split horizons as well for extra algorithm stability.

Poison Reverse Updates - for larger loops. Increased routing metrics usu. means a routing loop. Poison
reverse removes the problem route (via assigning very high metric) and places it in hold-down.

IGRP timers
-To control performance:

Update timers - frequency of routing update messages (default: 90 sec.)


Invalid timers - how long a router should wait before declaring a route invalid because it didn’t receive an
update about it.
Hold down timers - Specify the hold-down period. (default: 3x update timer + 10 sec)
Flush timers - time before should pass before a route should be flushed from routing table.

IGRP routes

IGRP advertises 3 types of routes

Copyright © 2004 cisconotes 23


Interior - routes between subnets. If network isn’t subnetted, it’s not advertised by IGRP.
System - routes to networks within an autonomous system. Obtained from directly connected interfaces,
other IGRP routes, or access servers. Does not include subnet information.
Exterior - routes to networks outside the autonomous system. Considered when using a gateway of last
resort. Gateway of last resort is chosen from list of exterior routes that IGRP provides.

Administrative distance or metric can be used to represent trust-worthiness of the routing update source
Each type of route and routing protocol is assigned an administrative distance. The higher value, the less
trusted the source.

Default distance
Directly Connected Interface 0
Static Route 1
Enhanced IGRP Summary Route 5
External BGP 20
Internal Enhance IGRP 90
IGRP 100
OSPF 110
IS-IS 115
RIP 120
EGP 140
Internal BGP 200
Unknown 255

Enhanced Interior Gateway Routing Protocol (EIGRP)


-Most like IGRP, but has more features.
-Proprietary Cisco.
-Combines advantages of link-state routing protocols with distance-vector protocols (balanced hybrid).
-Uses distance vectors for best path determination, resembles link state because it uses topology changes to
trigger routing database updates.

EIGRP includes
-Fast convergence
-Variable-length subnet masks
-Partial-bounded updates
-Multiprotocol support: IP/IPX/Appletalk

-EIGRP does not make periodic updates like RIP.


-When first communicates receives complete routing table; then only sends routing changes (partial
updates)… it only sends these changes to neighboring routers that would be affected by the changes.
Results in improved bandwidth for EIGRP.

Open Shortest Path First (OSPF)


-OSPF is based on link-state algorithms (known as a link state routing protocol). Deployed within an
autonomous system (share a routing protocol). OSPF routers retain its own topology database of it’s AS.

More flexible system with the following advantages


-Administrators are free to assign costs to a particular link. Total cost doesn’t have to have a limit. Upper
metric limit is 65,535. Each router creates link-state database tree representing the network with itself as
the tree’s root, it’s very good for choosing the most direct path. If more than one route of equal cost, OSPF
routers can balance the load between each equal cost route.
-Link-state routing advertisements are broadcasted much less often (only when change detected)
-Link state routing update packets can carry information for more than one router (sent only to adjacent
routers). These routers are chosen to swap the information.
-Can receive, send routes to other AS (even though OSPF is an interior routing protocol).

Copyright © 2004 cisconotes 24


-Type of service routing (TOS)
-Support for one or more metrics
-Variable-length subnet masks (VLSMs)

Exterior Routing Protocols


-Designed to communicate between AS’s (more complex than interior routing protocols due to more
information about greater number of routers).

Exterior Gateway Protocol (EGP)


-Dynamic routing protocol with a simple design… Routing updates specify only certain networks are
accessible through specific routers. EGP does not use metrics like interior routing protocols, so it can’t
detect or correct routing loops.
-EGP is a distance-vector protocol that allows ASs to communicate through a core routing network. These
AS’s have only one connection to the core, and connect to each other only through the core.

EGP Functions
-EGP routers establish a set of neighbors which they share accessibility info.
-EGP routers send polls to their neighbors to see if they are alive
-EGP routers send update messages containing accessibility information on the network within their
autonomous system.

EGP message types


-Neighbor acquisition messages to test if neighbors are alive
-Neighbor reachability messages to determine when a neighbor is down
-Poll messages to acquire accessibility information about the networks on which remote hosts reside.
-Error messages to identify various error conditions.

Border Gateway Protocol (BGP)


-BGP is an inter-autonomous system protocol created for use on the Internet. BGP can be used between
and within AS’s, and can determine routing loops. BGP neighbors can communicate only if on same
physical network. Comm. between routers in same AS helps determine connection point to or from
external AS’s.
-BGP doesn’t use a core like EGP: all AS’s can interconnect to all other AS’s.
When BGP comes online it receives complete BGP routing table from neighbors.
If change occurs, updates are sent incrementally.
BGP metric: degree of preference of a path… based on AS count, type of link, others. Usu. assigned by
administrator. via configuration files.

BGP message types


-Keeps track of all possible routes, but advertises the best route in update messages.
-Update messages have network number and an AS system path (path though multiple AS’s).
-This means BGP lists the route (not just next gateway) to destination; prevents counting to infinity and
slow convergence (found in other distance-vector protocols)

Other message types


-Open: first message sent after transport protocol connection is established
-Notification: sent when error is detected
-Keepalive: sent often enough to keep the hold timer from expiring.

Configuring Routing
Configuring RIP
sh ip route
-In config mode: Remove static entries: no ip router 172.16.50.0

Copyright © 2004 cisconotes 25


-If remotely: telnet into remote routers and remove static routes (without them can’t telnet directly to
router, will have to telnet from one router to neighbor router and then so on).
RIP accepts only major network (class A, B, C).

In configuration mode:
router rip
(config-router)#network 172.16.0.0
^Z

sh ip route should now show R entries (RIP).

Passive RIP – used to disable RIP advertisements out a particular interface (usu. The interface to the
internet)

config t
(config)# router rip
(config-router)# network 172.16.0.0
(config-router)# passive serial 1 Å new command
(config-router)# exit

-Opposite of passive RIP (will send RIP broadcasts across a wan link (non-broadcast network));
This sends rip broadcasts to the router specified in neighbor command

config t
(config)# router rip
(config-router)# network 172.16.0.0
(config-router)# neighbor 172.18.3.10 Å new command
(config-router)# exit

Monitoring RIP
sh ip route

[120/1] [admin distance/ metric]


admin distance:
RIP 120
OSPF 110
IGRP 100
EIGRP 90
STATIC 1
DIRECTLY CONNECTED 0

-It is used if two routing protocols advertise the same route to the same router.
-Will show time since last update, i.e. 00:00:06

sh ip route shows rip only

sh ip protocol : RIP timers, network for which RIP is assigned, and routing information sources (“Gateway
Information Sources”)

sh ip interface lots of statistics and interface configuration.

sh ip int [ethernet null serial brief] individual interfaces i.e.: sh ip int s0

debug ip rip shows the routing updates as they’re send/rec’d undebug rip

Copyright © 2004 cisconotes 26


trace 172.16.15.2 path packet takes to get to the final destination, shows each router’s accepting interface’s
IP, then the destination host’s IP.

Configuring IGRP
(config)#no router rip

(config)#router igrp 10 Å AS number, routers only comm. within an AS


(config-router)#network 172.16.0.0

-IGRP includes delay, bandwidth, reliability and load in update packets (even MTU), all to consider best
route or for load balancing. The IGRP packet is sent every 90 seconds.

Monitoring IGRP
-(same commands as RIP)

sh ip route igrp.. same [admin distance/cost]


-Routing Protocol is “igrp 10”
-Sending updates every 90 seconds, next due in 65 seconds (same as RIP, times usually different)

bandwidth command used on interfaces, IGRP uses this bandwidth value… default is 1544 Kbit (T1).

sh ip interface how an int. is configured & display status and global parameters for an interface.

debug ip igrp [events transactions]


undebug ip igrp [events transactions]

-Events is terse: displays IP address /serial sending or receiving update, and number of interior, system and
exterior in update
-Transactions is more verbose: includes the actual subnets sent / received in update with metric and
neighbor number.

Router Configuration
IOS Sources
-IOS source varies depending on hardware.
-Default is flash memory (EEPROM). TFTP & fallback routine are alternates.

copy flash tftp backup copy stored in TFTP server, can be used to verify same as original file.

-Will ask for address or name of remote host, then source, destination file name.

copy tftp flash

-Rarely enough room in flash for more than one version, so it will ask if ok to erase.

show flash verifies size matches original file. (size of files and amount of free flash)

delete allows removal of particular files.

delete does not remove flash files on external flash cards, only marks as deleted. Use squeeze to complete.

Fallback and Default Location of IOS

Copyright © 2004 cisconotes 27


boot system flash ios_filename
boot system TFTP ios_filename TFTP_address
boot system rom

-Those three commands above will be tried in sequence to find the IOS.

-Default location set by Configuration Register, usually set for first file in flash.

boot system [filename flash mop rcp rom TFTP]


boot [bootstrap buffersize host network system]

host – router-specific configuration file


network – network-wide configuration file

ROM typically lacks protocols, features, and configurations of full Cisco IOS s/w.
show version displays IOS version. Use this instead of show flash if more than one IOS file is stored.
Also shows router uptime and how was restarted, and where loaded from, hw config info, and configuration
register.

copy running-config tftp


Remote host []? 172.16.10.1
Name of configuration file to write [routerc-config]? Return
Write file routerc-config on host 172.16.10.1?
[confirm] return
Building configuration…
OK

copy TFTP run


Host or network configuration file [host]? Return
Address of remote host [255.255.255.255]? 172.16.10.1
Name of configuration file []? Detroit-config return
Configure using detroit-config from 172.16.10.1? [config] return
Loading detroit-config .. from 172.16.10.1 (via Ehernet0): ~
[ok – 717/32723 bytes]

Cisco AutoInstall Procedure


Used to setup routers remotely, great when no MIS staff avail in area.
Router acts as BootP workstation & connects to BootP or RARP server.

-This router gives the router it’s ip address and uses helper address to forward UDP broadcasts—TFTP
broadcast in this case, for a TFTP server. Also can be used to forward DHCP requests to a DHCP server.

-The new router will then look for a DNS (or TFTP) server to resolve it’s ip address to it’s host name.
-If from TFTP server, server sends “network-config” (applies to all routers) file: contains new hostname.
-Router requests “hostname-config” (applies to particular router), if not available, “router-config” – need to
telnet and make remaining changes manually.
-The file will then be downloaded from the TFTP server & loaded into running configuration.

IOS commands
sh proc

Shows active processes. CPU utilization for five seconds, one minute, and five minutes.
PID – Id # of each process
Q – queue (high, medium, low) priority
TY – is a status of the process.

Copyright © 2004 cisconotes 28


PC – is the program counter
Runtime – CPU time in milliseconds for the process
Invoked – total time the processes has been invoked
uSecs – CPU time in microseconds for each process invocation.
Stacks – low watermark/total stack space available in bytes.
TTY – tells which terminal controls the process.
Process – name of the process

sh mem
How management system allocated memory for different purposes

Head Total(b) Used(b) Free(b) Lowest (b) Largest (b)


Processor
I/O

Also can make a block by block listing of memory use: Address, bytes, prev, next, ref, prev F, next F, What

sh stacks – monitors the stack use of processes and interrupt routines, if reboot was result of crash has
reason

sh buff – reveals size of small, middle, big, very big, large and huge buffers.

Buffer elements:
500 in free list (500 max allowed)
726 hits, 0 misses, 0 created

Public buffer pools:


small buffers, 104 bytes (total 50, permanent 50)
50 in free list (20 min, 150 max allowed)
357 hits, 0 misses, 0 trims, 0 created
0 failures (0 no memory)

show running-config

show startup-config

sh int – shows interfaces and their status., h/w address, network address. sh int e0 – shows particular
interface Ethernet #0.

sh prot - which protocols are configured on the router.. will list each interface and it’s IP address (or other
protocol)

Global values:
Internet Protocol routing is enabled
Ethernet0 is up, line protocol is down
Internet address is 172.16.30.1 255.255.255.0

sh ip prot
Routing protocol is ‘rip’
Sending updates every 30 seconds, next due in 22 seconds
Invalid after 180 seconds, hold down 180, flushed after 240
Outgoing update filer list for all interfaces is not set
Incoming update filter list for all interfaces is not set
Redistributing: rip
Routing for Networks:

Copyright © 2004 cisconotes 29


172.16.0.0
Routing Information Sources:
Gateway Distance Last Update
172.16.40.1 120 0:00:02
Distance: (default is 120)

Accessing Other Routers


-SNAP on Data Link layer is used by devices running different network layer protocols (all LANs, most
`s).

-CDP starts by default on v10.3 and later, discovers neighboring Cisco routers running CDP via Data Link
broadcast.

-CDP caches information gained, which includes information about network layer protocols. Old CDP
information is discarded.

sh cdp int – interface information and encapsulation. This is the default encapsulation for the interface.
Also shows timers for CDP.

Ethernet 0 is up, line protocol is up, encapsulation is ARPA

sh cdp entry * - shows CDP information from all routers. replace * with a router name for just a specific
router

Displays router’s name, entry’s IP address, platform [i.e. Cisco 2500], Capabilities: Router Interface:
Serial0, port ID (outgoing port): Serial0
Holdtime: 130 sec

Version

Cisco Internetwork Operation System Software


IOS v11.0(18)….

sh cdp neighbors – reveals information exchanged among neighbors.

Capability Codes: R – Router, T – Trans Bridge, B- Source route Bridge …

Device ID Local Intrfce Holdtme Capability Platform Port ID


Router C Ser 1 158 R 2500 Ser 0

Device ID – name of neighbor router


Local interface – the interface on which this neighbor is heard
Holdtime – decremental hold time in seconds
Capability – router’s capability code R for router, S for switch, etc.
Platform – which type of device
Port ID – the interface of the remote neighbor router

sh cdp neighbor detail

Same information as above, plus information from sh cdp entry.

sh cdp traffic amount of CDP packets sent and received among neighbors, plus packet errors

Virtual Terminal Connections (Telnet)


-Up to 5 simultaneous telnet sessions on Cisco.

Copyright © 2004 cisconotes 30


RouterB#telnet 172.16.50.1
Trying 172.16.50.1 … Open

User Access Verification

Password:
RouterC>exit
[ Connection to 172.16.50.1 closed by foreign host]

To return to originating router without terminating the session to remote router: CTRL+SHIFT+6 then X
Then to see sessions:
RouterC# sh sessions

Conn Host Address Byte Idle Conn Name


* 1 routerb 172.16.40.1 0 0 routerb
RouterC# 1
[ Resuming connection 1 to routerb … ] (return)
RouterB>

Type end or disconnect to terminate the sessions, or add unlimited sessions.

Hostnames
-Router can be configured to resolve hostnames like a hosts file in Unix --OR-- use a DNS server.

config t
ip host ?
WORD Name of host

RouterB (config)# ip host RouterC ?


<0-65535> Default telnet port number
A.B.C.D Host IP address (maximum of 8)

ip host RouterC 172.16.40.2


^Z

routerc
Trying RouterC (172.16.40.2)…Open

User Access Verification

Password:
RouterC>exit

[Connection to routerc closed…

sh hosts –displays all hostnames and their IP addresses

Hosts Flags Age Type Address(es)


RouterC (perm, OK) 0 IP 172.16.40.2
RouterA (perm, OK) 0 IP 172.16.20.1

Flags – how name was learned


Perm – manually configured
Temp – if resolved by DNS

Copyright © 2004 cisconotes 31


OK – it’s current
EX – entry has timed out
Age – time in hours the name has been resolved
Type – protocol
Address(es) logical address of the host.

ip domain-lookup enabled by default – router will try to broadcast for a DNS server to resolve the name.

no ip domain-lookup stops the router from long timeouts before failing to find a DNS server.

ip name-server 172.16.30.8 is used with IP domain-lookup so the router can lookup hostnames.

Basic Testing
-Follow layers of OSI model: Telnet – Application layer, ping / trace / show ip route – Network layer, show
interface – physical
-If can’t telnet, use ping to test physical network connectivity. Can be used with IPX, AppleTalk, Apollo,
VINES, and DECnet. PING is short for “Packet InterNet Groper”.

PING Results Defined


! receipt of echo reply
. time-out
U destination unreachable
C congested experience packet
I ping interrupted (CTRL-Shift-6 X)
? packet type unknown
& packet time-to-live exceeded

Extended ping: more options, can use IP, AppleTalk, IPX, & only avail. in privileged mode.

RouterC# ping
Protocol [ip]:
Target IP address: 172.16.10.2
Repeat count [5]:
Datagram size [100]:
Timeout in seconds [2]:
Extended commands [n]: y
Source address or interface: 172.16.40.2
Type of service [0]:
Set DF bit in IP header? [no]: y
Validate reply data? [no]: y
Data patter [0Xabcd]:
Loose, Strict, Record, Timestamp, Verbose [none[: v
Seep range of sizes [n]:

Testing with the Trace Command


-Used to discover routes to remote destinations. Similar to ping, but, Trace command takes advantage of
error messages when a packet exceeds its TTL.

-Trace command starts out by sending a TTL of 1 so 1st router will receive packet and send back an error.
Keeps sending out packets with incrementing TTL until the complete path, distance, and time is measured
between each router.

trace ip 172.16.40.2

Trace Responses

Copyright © 2004 cisconotes 32


!H The router didn’t forward the probe because of an access list.
P The protocol was unreachable
N The network was unreachable
U The port was unreachable
* There was a time-out

ICMP sends back an error to sending router with TTL of 125.

Other Testing Commands


sh int s0
Serial0 is up, line protocol is up

-Carrier Detect brings up the serial port (physical layer okay). Keepalives bring up the line protocol (data-
link framing okay – means have correct frame type and are communicating end to end).

-int up, line prot down: problem with connection or clocking.


-int down, line prot down: interface problem.
-int admin down, line prot down: interface disabled.

Note: Link light will go out when loopback is not activated on Ethernet jack on the other end of the cable.
This could also be an indication that the Ethernet wiring to the Ethernet jack on the other end is bad. In
these two cases, show int command will show: int down, line prot down.

sh int shows real-time statistics of an interface: input errors – problem with link or if equipment is faulty.
clear counters

-Debug command – see earlier section where it is used for tracking RIP and IGRP. Track network
problems, and discover which protocol messages are being sent / recd.

NOTE: debugging live network could cause severe slowdown. Do not leave it on- diagnose then turn off.

RouterC# debug all


This may severely impact network performance. Continue?
[confirm]
All possible debugging has been turned on
RouterC#
Serial0: HDLC myseq 2139, mineseen 2139, yourseen 2139, line up
CDP-PA: Packet received from RouterB on interface Serial0
** Entry found in cache **
RIP: sending update to 255.255.255.255 via Ethernet0
(172.16.50.1)
subnet 172.16.40.0,. metric 1
subnet 172.16.20.0, metric 2
subnet 172.16.10.0, metric 3
RIP: Update contains 3 routes
RIP: Update queued

undebug all

debug ?

RouterC#config t
Enter configuration commands….
RouterC(config)#logging 172.16.10.1 – logs messages to a log file on a Unix or NT host instead of
console terminal.
Syslog logging: enabled (0 messages dropped, 0 flushes, 0 overruns)

Copyright © 2004 cisconotes 33


Console logging: level debugging, 39 messages logged
Monitor logging: level debugging, 0 messages logged
Trap logging: level informational, 16 message lines logged

Logging to 172.16.10.1, 1 message lines logged


RouterC#

IPX/SPX Protocol Suite


-Prior to Novell Netware 5.0, IPX was default protocol for Novell. Novell was most popular NOS during
80’s and early 90s. Novell still supports IPX.

-Internetwork Packet Exchange used since early 80s, similar to XNS, likeness to TCP/IP.
IPX is a family of protocols.

-IPX developed for performance rather than strict compliance to existing standards or models.
IPX doesn’t map directly to OSI model, but still functions in layers.

IPX (Internetwork Packet eXchange)


-Performs functions at layers 3 and 4 of OSI model.
-Assignment of IPX addresses
-Governs packet delivery, and makes routing decisions based on RIP or NLSP.
-Connectionless like TCP/IP’s UDP (no ACKs from destination node)
-IPX communicates with upper-layer protocols via SOCKETS. These are similar to TCP/IP ports (used to
address multiple independent apps on the same machine)

SPX (Sequence Packet eXchange)


-Adds connection oriented communications to the connectionless IPX. It ensures data delivery for upper
layers. SPX does this via virtual circuits or connections, with each connection having a connection ID in
SPX header.

RIP (Routing Information Protocol)


-Distance-vector protocol used to discover IPX routes through the internetworks. Uses ticks (1/18 th of a
second) and hop count (number of routers between nodes).

SAP (Service Advertising Protocol)


-Used to advertise and request services (for servers and clients, respectively)

NLSP (Netware Link Services Protocol)


-Advanced link-state routing protocol developed by Novell. Replacement for RIP and SAP.

NCP (NetWare Core Protocol)


-Provides clients with access to server resources (file, printing, synchronization, and security)
-All these protocols: routing, connection/connectionless, & has application protocols: means IPX capable
of supporting large intranets running many applications.
-Netware follows strict client-server model – must be either client or server.
-Netware servers provide file, print, message, application, and database svcs.
-All clients depend on Netware servers to locate all network resources.
-All Netware servers build a SAP table of all network resources it is aware of.
-Clients issue IPX broadcast called GNS (GetNearestServer request)
-Servers receiving GNS request check SAP tables to locate a Netware server that matches the specific
request – it responds to the client with another GNS (GetNearestServer response).
-If no servers hearing the GNS request have the requested service or know of a server that does, no
response is created.

Copyright © 2004 cisconotes 34


Cisco routers create SAP tables too
-They can respond to GNS requests just like servers.
They don’t give the service, but respond to requests with the location of a server that does.
-GNS response can come from local/remote NetWare server, and Cisco router. Local servers should
respond first, otherwise Cisco router can respond (quicker than remote Netware servers, prevents WAN
traffic).

Server-Sever Communication
-More complicated, & maintains tables of avail network resources.
-Each server must be able to locate any resource.
-Servers exchange 2 types of info, using 2 protocols:
-SAP (Service Advertising Protocol)
-RIP (Routing Information Protocol)

-RIP in IPX is a completely separate routing protocol than RIP in TCP/IP, it’s not the same routing
protocol.

-SAP broadcast sent every 60 seconds


-SAP broadcast includes all services learned from other servers not just server itself.
-All servers receiving the broadcast incorporate it into their own SAP tables
-All servers eventually become aware of all services (can respond to GNS requests)

-By default, SAP broadcast won’t cross a CISCO router.


-Cisco router catalogs all SAPs heard on any IPX enabled interface in its SAP table
-Cisco router then broadcasts the whole table (summarized) from each IPX interface at 60 second int. by
default.

-RIP information is exchanged similarly.


-Servers build routing tables without entries for directly connected networks, then broadcast through all
IPX interfaces.
-Other servers receive those updates then broadcast that information to all IPX interfaces.
-RIP information is communicated until all servers and routers know of the internetworks routes. (like
SAP)
-RIP default broadcast interval is 60 secs.

IPX Addressing
-10 bytes (80 bits)
-hierarchical: network / node portions 4 byte network (8 hex) / 6 byte node (12 hex):
00007c80.0000.8609.33e9
-network portion assigned by administrators and must be unique on IPX internetwork
-node addresses automatically assigned (usu. MAC address used); no need for DHCP or manual
configuration
-because MAC is used for client address, no need for TCP/IP ARP equivalent in IPX.
-Customary to drop leading 0’s in network portion: 7c80
-Commonly node is divided into 3 sections of four hex digits.
-Networks running both TCP/IP and IPX administrators freq use TCP/IP’s 8 hex digits as the IPX network
address on that segment.

Encapsulation - Process of taking packets from upper layer protocols and building frames around them to
transmit. In an IPX network, encapsulation is when IPX datagrams have layer 2 frames built around them.

-Netware has four different frame types, each incompatible with the other (but a router can make
communication possible but only across the router).

Ethernet_802.3 Default to Netware 3.11

Copyright © 2004 cisconotes 35


Etherent_802.2 Default since Netware 3.12
Ethernet_II Supports TCP/IP and IPX
Ethernet_SNAP AppleTalk, IPX, and TCP/IP

Note: segmentation can refer to Ethernet (token ring, etc.) segments or cable segments.

-Can create three virtual networks on the same Ethernet segment by using three different frame types; This
requires 3 unique IPX network addresses for each frame type. Communication between same frame type is
possible.. A router is needed to facilitate communication between different frame types. If configuring a
router, get the frame type and IPX network address for each segment, type config at the server console.

Enabling IPX on Cisco Routers


RouterA# config t
RouterA(config)# ipx routing –turn on IPX routing
RouterA(config)# ^Z
%SYS-5-CONIFG_I: Configured…
RouterA# sh prot
Global values
Internet Protocol routing is enabled
IPX routing is enabled
Ethernet0 is up, line protocol is up
Internet address is 172.16.20.1/24
Serial0 is up, line protocol is up
Internet address is…..

Enabling IPX on Individual Interfaces


-Enter interface configuration mode, then issue the following command:
ipx network number [encapsulation encapsulation-type] [secondary]

-Secondary is optional secondary encapsulation on same interface

Novell IPX frame types


-Used in the above command's encapsulation-type
CISCO KEYWORD
Ethernet_802.3 novell-ether (default)
Ethernet_802.2 sap
Ethernet_II arpa
Ethernet_snap snap

Token Ring sap (default)


Token-Ring_snapsnap

Fddi_snap snap (default)


Fddi_802.2 sap
Fddi_raw novell-fddi

(hint: 802.2 = SAP - Service Advertising Protocol)

config t
int e0
ipx network 2100
int s0
ipx network 2200

-Make sure Novell servers are running the same network number as router’s interface.

Copyright © 2004 cisconotes 36


Adding Secondary Addresses
ipx s0
ipx network 3200 encap hdlc sec
exit
sh prot s0
Serial0 is up, line protocol is up
Internet address is 172.16.20.1/24
IPX address is 2200.0000.0c8d.5c9d (HDLC)
IPX address is 3200.0000.0c8d.5c9d (HDLC)

Note: This method is not supported in future IOS releases.

-Better way: use subinterfaces— allows creation of virtual interfaces: can use for sec IP, IPX, etc.
interface ethernet slot/port.number from e0.0 to e0.4billion.. note: each one takes memory!
int e0.100
ipx network 2300 encap sap
^Z
%SYS-5-CONFIG_I: Config…
sh prot e0
Ethernet0 is up, line protocol is up
Internet address is 172.16.10.1/24
IPX address is 2100.0000.0c8d.5c9d
sh prot e0.100
Ethernet0.100 is up, line protocol is up
IPX address is 2300.0000.0c8d.5c9d

Monitoring and Troubleshooting IPX on Cisco Routers


sho ipx servers (like display servers command in Netware) displays SAP table in Cisco router.
Make sure remote servers entries are in the SAP table, if not check IPX address, and encap settings.

Codes: S – Static, P – Periodic, E – EIGRP, N – NLSP, H - Holddown, + = detail

Table ordering is based on routing and server info

Type Name Net Address Port Route Hops Itf


P 4 BORDER1 350ED6D2.0000.0000.0001:0451 2/01 1 Et0
P 4 BORDER3 12DB8494.0000.0000.0001:0451 2/01 1 Et0

sh ipx route

Codes: C – Connected primary network, c – Connected secondary network, S – Static, F- Floating static, L
– Local (internal), W – IPXWAN, R – RIP, E – EIGRP, N – NLSP, X- External, A – Aggregate, s –
seconds, u – uses

6 Total IPX routes. Up to 1 parallel paths and 16 hops allowed. *SEE BELOW

No default route known.

C 2100 (NOVELL-ETHER), Et0


C 2200 (HDLC), Se0
C 2300 (SAP), Et0.100
c 3200 (HDLC), Se0
R 4100 [07/01] via 2200.00e0.1ea9.c418, 13s, Se0

Copyright © 2004 cisconotes 37


R 5200 [13/02] via 2200.00e0.1ea9.c418, 13s, Se0

- should see internal IPX network along with external IPX network on this list. Use configuration
screen at server console also.

Note: With parallel IPX paths, by default CISCO IOS does not learn alternative equal-cost paths.

ipx maximum paths 2 (up to 512). Allows parallel paths.

Cisco will do per-packet load-sharing by default over parallel paths. Packets are sent round-robin on equal
cost lines.

IPX per-host-load-share to ensure certain destinations go over only one particular interface.

sh ipx traffic –summary of the number and type of IPX packets sent/recd, shows IPX RIP & SAP updates.

System traffic for 0.0000.0000.0001 System-Name: RouterA


Rcvd: 15 total, 0 format errors, 0 checksum errors, 0 bad
Hop count, 0 packets pitched, 15 local destination, 0 multicast

Bcast: 10 received, 249 sent


Sent: 255 generated, 0 forwarded
0 encapsulation failed, 0 no route
SAP: 1 SAP requests, 0 SAP replies, 0 servers
0 SAP Nearest Name requests, 0 replies
0 SAP General Name requests, 0 replies
0 SAP advertisements received, 0 sent
0 SAP flash updates sent, 0 SAP format errors
RIP: 1 RIP requests, 0 RIP replies, 6 routes
8 RIP advertisements received, 230 sent
12 RIP flash updates sent, 0 RIP format errors
Echo: Rcvd 0 requests, 5 replies
Sent 5 requests, 0 replies
0 unknown: 0 no socket, 0 filtered, 0 no helper
0 SAPs throttled, free NDB len 0
Watchdog:
0 packets received, 0 replies spoofed
Queue lengths:
IPX input: 0, SAP 0, RIP 0, GNS 0
SAP throttling length: 0/(no limit), 0 nets pending lost route reply
--More--

sh ipx int e0

Ethernet0 is up, line protocol is up


IPX address is 2100.0000.0c8d.5c9d, NOVELL-ETHER [up]
Delay of this IPX network, in ticks is 1 throughput 0 link delay 0
IPXWAN processing not enabled on this interface.
IPX SAP update interval is 1 minute(s)
IPX type 20 propagation packet forwarding is disabled
Incoming access list is not set
Outgoing..
IPX helper access list is not set
SAP GNS processing enabled, delay 0ms, output filter list is not set
SAP Input filter list is not set
SAP Output filter list is not set

Copyright © 2004 cisconotes 38


SAP Router filter list is not set
Input filter list is not set.
Output…
Router …
Netbios Input host access list…
Netbios Input bytes access list
Netbios Output
Netbios Output

debug ipx – real time debugging

debug ipx routing [or specific routing protocol, or other]

debug ipx routing act – activity

IPX routing debugging is on


RouterA#
IPXRIP: update from 2200.00e0.1ea9.c418
5200 in 2 hops, delay 13
4100 in 1 hops, delay 7
IPXRIP: posting full update to 2100.ffff.ffff.ffff via Ethernet0 (broadcast)
IPXRIP: src=2100.0000.0c8d.5c9d, dst=2100.ffff.ffff.ffff, packet sent
Network 5200, hops 3, delay 14
Network 4100, hops 2, delay 8
Network 2300, hops 1, delay 2
Network 3200, hops 1, delay 2
Network 2200, hops 1, delay 2
….
undebug ipx routing act
IPX routing debugging is off

Extended Ping
RouterC# sh prot e0
Ethernet0 is up, line protocol is up
Internet address is 172.16.50.1/24
IPX address is 5200.0000.0c3f.1d86
RouterA# ping
Protocol [ip]: ipx
Target IPX address: 5200.0000.0c3f.1d86
Repeat count [5]:
Datagram size[100]:
Timeout in seconds [2]:
Verbose [n]:
Novell Standard Echo [n]: y
Type escape sequence to abort.
Sending 4, 100-byte IPX Novell Echoes to
5200.0000.0c3f.1d86, timeout is 2 seconds:!!!!!
Success rate is 100 percent (5/5), roundtrip min/avg/max = 4/7/12 ms

Access Lists & Traffic Management


-Access lists are packet filter lists which can act on, categorize, and compare packets. Can be applied to
inbound or outbound traffic on any interface – the router will then have to analyze every packet crossing
the interface in the specified direction (and take action).
-Packets are always compared to each line from top to bottom in the access list
-The comparison stops immediately when a match occurs.

Copyright © 2004 cisconotes 39


-An implicit deny is at the end of the access list.

Standard IP Access Lists


-Configure access list, then apply to an interface.
-Standard IP access lists: SOURCE address considered ONLY

access-list {number} {permit or deny} {source address}

1-99 Standard access list


100-199 IP extended access list
800-899 IPX standard access list
900-999 IPX extended access list
1000-1099 IPX SAP access list

config t
int e0
ip access-group 10 out – this applies the list to an interface on outgoing traffic .
^Z

-Only one access list per protocol is allowed to be applied on inbound (or outbound) traffic.

Wildcard Masking
config t
access-list 11 permit 172.16.50.2 0.0.0.0
access-list 11 permit 172.16.30.0 0.0.0.255 (binary 1’s in mask means any)
int e0
ip access-group 11 out

Note: common practice is to enter a new list, apply it, then remove old one.. then later, cycle back.

Extended IP Access Lists


Acts on source, destination, IP protocol (TCP, UDP, ICMP, etc), and PORT (WWW, DNS, FTP, etc.)

access-list {number} {permit | deny} {protocol} {source} {destination} {port}


^ ^ ^
access-list 110 permit tcp host 172.16.50.2 host 172.16.10.2 eq 8080
access-list 110 permit tcp 172.16.30.0 0.0.0.255 host 172.16.10.2 eq 8080
access-list 110 permit tcp any any eq www

-The “host” keyword is equivalent to 0.0.0.0 wildcard mask.


-The “any” keyword is equivalent to 0.0.0.0 255.255.255.255

Monitoring Extended IP Access Lists

show access-list – shows all access lists running


Extended IP access list 110
Permit tcp host 172.16.50.2 host 172.16.10.2 eq 8080 (34 matches)
Permit tcp 172.16.30.0 0.0.0.255 host 172.16.10.2 eq 8080 (11 matches)
Permit tcp any any eq www (33 matches)

(apparently the book says that these counters are only on extended access lists)

-Administrators typically move the most used lines up in the list for efficiency, be careful not to change
functionality by doing so.

Copyright © 2004 cisconotes 40


access-list 110 deny ip any any log – by default will log to the console:

%SEC-6-IPACCESSLOGP: list 110 denied tcp 172.16.30.2(2024) -> 172.16.10.2(21), 1 packet

Included in the log:


-access list number, source address, source port, destination address, destination port, number of packets.

-Also when using show access-list:


deny ip any log (4 matches)…
-All log information can be redirected to syslog server and stored.

show ip access-list only IP access lists.

clear access-list counter (s)? clears counters for show access-list commands.

show ip interface e0 – shows IP settings for interface & access list applied to it.
show running config – shows router’s current configuration & access list applied to each interface.

Standard IPX Access Lists


-Allow or deny packets based on source and destination IPX addresses.

access-list [number] [permit/deny] [source] [destination]

access-list 810 permit 30 10


access-list 810 deny 50 10
int e0
ipx access-group 810 out

(note: no wildcard masking is necessary to specify an entire IPX network)

-To allow all IPX nets except 50 you would: deny 50 10 then: permit -1 –1
-The -1 means any IPX network (like the any keyword for IP access lists)

Extended IPX Access Lists


-Allow or deny based on source, destination, IPX protocol (SAP, SPX, etc.), IPX socket

access-list {number} {permit | deny} {protocol} {source} {socket} {destination} {socket}


^ ^ ^
(hint: moving from std to extended – addition of protocol and socket options)

access-list 910 deny –1 50 0 10 0


access-list 910 permit –1 –1 0 –1 0

0 socket means any IPX socket

IPX SAP Filters


-Controlling SAPs; That controls access to IPX devices.

access-list {number} {permit/deny} {source} {service type}

Copyright © 2004 cisconotes 41


access-list 1010 permit 11.0000.0000.0001 0 (only addresses starting with 11 will be seen by outside
world)
int e0
ipx input-sap-filter 1010
^Z

-0 service means any service type


-Above command allows packets with addresses starting with 11 to pass the Ethernet interface and be
included in SAP updates across the network (added to SAP table).

VLANs (Virtual LANs)


-Creates smaller Ethernet broadcast domain (Virtual LAN) out of one big Ethernet broadcast domain.
-Creating a VLAN results in creating a smaller broadcast domain by assigning one or more switch ports to
a broadcast domain (broadcast domain is not necessarily equal to the IP subnet).
-Each VLAN is a separate broadcast domain.
-VLANs act on layer 2 (frames) only.

For exam, know:


-7 reasons to create a VLAN(s).
-VLAN frame filtering
-VLAN frame tagging
-What switches do to create VLAN(s)
-How to use hubs with switches.
-5 components when using VLANs
-Cisco VLAN design
-Different VLAN technologies
-Token Ring VLANs
-VLAN types: end-to-end, local VLANs

VLAN - logical group (location ,department, security, layer 3 protocol, application, etc.) of resources and
users to switch ports that is configured by an administrator.
-Behaves like a subnet because it defines the limits of an broadcast domain (Ethernet broadcast domain,
Token Ring broadcast domain, etc.)
-Lower costs, easier administration, smaller broadcast domain, separation of traffic and services,
security, flexibility, scalability.

VLAN Types
end-to-end: most traffic remains on VLAN (80/20 rule). VLANs by job function or department. Users do
not have to be in the same switch block.

local VLANs: based on geographic locations (demarcation at a switch block boundary) 20/80 rule is
typically followed. These are used when the network is growing fast & is constantly changing.

VLANs mean Easier administration and lower costs


-VLANs great for mobile users, they will still be in the same VLAN no matter where they are, only need
either a quick reconfigure of the user’s switch port to make it part of their VLAN or enable auto join
VLAN on the user’s switch port.
-Network addresses remain the same.
-Router configuration remain the same.
-Maximum users = 1000.
-Cisco Recommended maximum users = 150-200.

Broadcasts
-Today, switched networks are often preferred over backbones, since they are often cheaper.
-Broadcasts make it through switches by default.

Copyright © 2004 cisconotes 42


-Broadcasts cannot make it through routers to other network segments by default.

flat network – physical network(s) with only one broadcast domain


Routers, and or RSMs (Route Switch Modules) are used in switched networks to have firewalls between
VLANs (this also stops broadcasts).

Security
-With backbone networks (instead of switch networks), security was provided by the router. This means
that each of the LAN segments connected to the router did not have security within the LAN segment itself.
Example: A network analyzer attached to a hub port on a LAN segment can see all the traffic going into the
hub.
-VLANs, by contrast, have control over each user, each port, MAC address, protocols, applications, & size
of broadcast domains.
-VLANs can be assigned on a per resource basis, allowing a network management station to alert
administrators of unauthorized access.
Inter
-VLAN security is done on the router or RSM.

Scalability & Flexibility


switch-fabric – multiple switches that are connected.
-Switch ports and/or users can be assigned to a VLAN(s).
-Assign users that belong to a VLAN, no matter where on the network they are.
-VLANs stop broadcast storms from getting across the entire network (they stay in the broadcast domain).
-Broadcast storms come from bad NICs, or rogue applications.
-Can break up big VLANs to smaller VLANs.
-Each VLAN gets a number assigned.
VLAN organizations
-Groups with the same interests.
-Routers and RSMs provide inter-VLAN communication.

Distribution of Traffic
-80/20 rule: 80% of traffic stays on local broadcast domain – when using bridges or VLANs.

Distribution of Network Services


-Servers can be trunked Æ prevents traffic from crossing VLANs.
-Attach a server to a switch port and can assign more than one VLAN or all VLANs (this is default) (note:
1005 VLANs is max).
-Cisco’s VlanDirector uses SNMP to configure and manage VLANs, and Cisco switches.

VLAN Features
-Static or Dynamic VLANs
-Frame Filtering
-Frame Tagging methods (for VLAN identification) supported by CISCO:
-ISL (Inter-Switch Link) for Ethernet/Fast Ethernet (Cisco proprietary technology)
-Special Ethernet NIC that understands/creates tagging
-Driver creates many logical cards
-Network addresses configurable for each VLAN
-Protocol analyzers can be used to examine frames on ISL trunk segments.
-Adds : 26 byte header, 4 byte trailer (CRC)
-Uses ASICs and processors to deal with 100Mbps frame tagging
-IEEE 802.1Q for Ethernet/Fast Ethernet
-Standardized frame tagging by IEEE
-Modifies frame (frame size kept same) :
-Tag Protocol Identifier (TPID) (4 byte)

Copyright © 2004 cisconotes 43


-Tag Control Information (TCI) (2 byte)
-Can be forwarded by non-802.1Q aware device (can cause leakage into other VLANs)
-Cisco 802.10 for FDDI
-ATM Forum’s LAN Emulation (LANE)
-Spanning Tree Protocol
-VLANs over backbone(s) such as ATM, FDDI, Fast Ethernet (connects different media type).
-VLAN trunks by default carry all VLANs (VLAN is identified by VLAN ID in frame).

VLANs Over Several Backbones


-VLAN Trunk Protocol (VTP) groups VLAN trunk ports between switches.
-VTP is an industry standard.
-VTP does not cause a lot of traffic.
-VTP works on Layer 2 and/or Layer 3 VLANs.
-VTP information is used by routers, switches, and servers.
-VTP can auto-configure switches.

-“fast backbone” means Fast Ethernet, ATM, or FDDI, these are transport protocols used for a
trunked link Æ port(s) assigned to many VLAN(s). Must have at least 100Mbps Ethernet for trunk port.

-Switches using VTP advertise their management domain, known VLANs with their boundary, & configure
revision number on their trunk ports.
-The connected VTP devices that have the same VTP domain name defines a VTP domain.
-A switch can only be assigned to 1 VTP domain

In a VTP domain:
-All VLANs are maintained by its servers and clients.
-VTP domain sets boundaries of all VLANs in it.
-Clients and servers send and receive information through trunks to other switches (belonging to other VTP
domain(s))

VTP servers store information:


-via TFTP or NVRAM (nonvolatile random-access memory).
-Change VTP servers’ information via VTP MIB (Management Information Base) or CLI (command-line
interface) Æ tells clients and servers that they will get traffic on their trunk ports (when VLANs are added
and/or advertised).
-VTP servers can allow switches to delete a VLAN and all ports assigned to the VLAN.

Configuring a New VLAN


-Do it on one device in management domain.
-All other VLAN devices learn this information via advertisement frames (sent to a multicast address).
-Advertisement frames rec’d by all neighboring devices (not normally forwarded).
-VTP sent on all trunk links (LANE, 802.1q, ISL)

Static VLANs
-Most common, most secure VLAN type
-Switch ports assigned to a VLAN by an administrator will stay that way until an administrator changes it.

Configure switch via:


-Network management software—i.e. Cisco Works for Switched Internetworks (NT or UNIX) via SNMP
-Console cable and using command line interface (CLI).
-Telnet not possible

Catalyst 5000 configuration example:


5000> (enable) set vlan 1 name ADMIN – VLAN 1 activated, and named “ADMIN” (optional)
Vlan 1 configuration successful

Copyright © 2004 cisconotes 44


5000> (enable) set vlan 1 3/1-3 – Card 3 (aka slot 3) ports 1 to 3 assigned to VLAN 1
VLAN 1 modified.
VLAN Mod/Ports
---- -------------------------
1 3/1-3
set vlan 1 3/1-3,2/1 is also a valid command.

Dynamic VLANs
-Decides which VLAN a node belongs to automatically using management software.
-Makes administration easy after initial configuration, and users can move easily.
-MAC address, various ports., and sometimes applications can be automatically assigned to a dynamic
VLAN
-MAC address – to – VLAN # & VLAN configuration database is created for VLAN management
software. If a node with one of these MAC addresses is attached to an unassigned VLAN port, the VLAN
management software will configure the switch port for the MAC address’ corresponding VLAN number
and VLAN configuration.
-VMPS (VLAN Management Policy Server is an example of a MAC address – to – VLAN # database.

VMPS (VLAN Management Policy Server)


-Stored on TFTP host.
-The database is downloaded when the switch s/w is started.
-UDP used to listen for client requests and it is also used for other communication.
-Explicit deny for any MAC address(s) is an option.
-If no match is found for a MAC address, the switch port is shut down

Dynamic Port VLAN Membership


-ONLY 1 VLAN can be assigned to a dynamic port (not a trunk port) at one time on Cisco Catalyst
switches.
-Dynamic port is isolated from the static VLAN when the link is activated.
-1ST MAC Source address sent from a new host is compared against VMPS database.
-If no match found the port is shut down if secure mode is enabled.
-If there’s a match, VMPS give the corresponding VLAN number for that port to be assigned.
-Dynamic VLAN not supported with 3-port Gigabit Ethernet module (WS-X5403)
-More than one host, each with different MAC addresses can be active on a dynamic port, as long as they
belong to the same VLAN.
-Switch port reverts to isolated state when link goes down.
-VMPS must be configured prior to configuring ports as dynamic.
-Reconfiguring a port from static to dynamic immediately connects the port to the VLAN
-VMPS checks after an interval to see if the MAC address or host is permitted to be on the dynamic port
-Must turn off security on a static port FIRST, before they can become a dynamic port.
-Must turn off trunking on a static port FIRST, before they can become a dynamic port.
-VTP management domain (management VLAN for VMPS clients) & VMPS server must be same.
-Any port that is configured as dynamic: Spanning-tree PortFast is enabled automatically by default.
-This prevents applications on hosts from getting bad configuration information, resulting in applications
entering loops or timing out.

Menu-Based VLAN Configuration


VLAN & VTP Configuration
-Normally the 2nd item after selecting the “Configuration” item from the main menu
-Contains the following 5 submenu choices:

Local VLAN Port Configuration


VLAN port assignments. Select [C]hange to change one. Select More for next 14 ports.

Copyright © 2004 cisconotes 45


VTP Administrative Configuration
-Domain name
-Domain mode:
-Client (changes only from remote devices)
-Server (changes only from local device)
-Transparent (lets VTP packets pass through).

VTP VLAN Configuration


Lists VLANs, & can change VTP assignment.

Local Preferred VLANs Configuration


Lists ALL VLANs.

Reassign Ports in Local VLAN


Moves an existing, configured Catalyst Stack to a pre-existing VTP administrative domain.

V – Virtual LAN Configuration


-At least 1 port should be assigned to VLAN #1
-Default = all port assigned to VLAN #1, and the management domain is VLAN #1.
-Cisco 2820 creates VLANs by using bridge groups.

N – Domain Name – sets a management domain name (before VLAN creation). Default = no domain
name. Switch can receive an advertisement with management domain name (and set it automatically that
way)

V – VTP Mode Control – [T] ransparent or [S] erver (default). Default Server mode, receives
advertisements on trunk port. Switch will change from VTP server mode to VTP client (transparent) mode
if advertisement with over 128 VLANs comes in.

F – VTP Pruning Mode – restrict flood traffic of VLAN to just switches that are members of the VLAN.
Each trunk has a configured list of pruning eligible VLANs.
A – Add VLAN – Adds VLAN to list of allowed VLANs for the trunk. Default=allow all VLANs on one
trunk.
Choose type 1 = Ethernet
N - set VLAN number
V - set VLAN name
I - set 802.10 SAID value (has to be in range displayed & can’t be duplicated)
M - set MTU size
T – set VLAN state to enabled (enabled/disabled)

M – Modify VLAN – change existing VLAN.


D – Delete VLAN – delete a functioning VLAN. Ports that were assigned to that VLAN go back to
VLAN#1.

T – Trunk Configuration – menu for Configuring VLAN trunks – carry multiple VLAN traffic between
switches and/or VLAN enabled routers. VLANs are able to get to multiple switches this way.
Choose trunk port A or B, enable trunking on that port via T then 1.

-VTP global configuration values (VTP advertisements via multicast address):


-VTP management domain name (defines boundary for a VLAN)
-VTP configuration revision number
-VLAN ID numbers (ISL, & 802.1Q). or ELAN names for ATM LANE. or SAID values for 802.10
(FDDI).
-Optional password (an MD-5 key)
-“subset” advertisements also include:

Copyright © 2004 cisconotes 46


-Type of VLAN (ATM, FDDI, Ethernet)
-VLAN configuration i.e. MTU (Maximum Transmission Unit) for the VLAN
-Frame format

-A switch can be configured in 3 different VTP modes:


-Server, Client, Transparent.

-Servers and Clients communicate via ALL VLAN trunks (including all types of trunks).
-Servers are responsible for all VLAN adds, changes, & deletions. They do this by multicast
advertisements to all other servers & clients.
-Server is default mode for Catalyst switches.
-Clients only accept changes from servers.
-Servers save VLAN information in NVRAM. Clients do not save VLAN information in NVRAM.
-Transparent mode switches do not belong to a management domain.
-Transparent mode switches forward VTP information to other switches.
-Transparent mode switches can do VLAN adds, changes, & deletions (but this is NOT SHARED with
other switches)

Note: clear config all clears switch configuration from NVRAM, but not the VTP information from
NVRAM ! use reset system or power cycle the switch instead (the revision number will also be set to 0
when you do this.)

When Connecting a new switch to existing network that uses VTP, you must clear VTP information first:
-Catalyst 5000: clear config all
-1900/2820: delete nvram
-2900XL: delete flash:config_file_name
-Then, power cycle the switch
-Then, configure VTP settings
-Finally, add it initially as a client.

-Use only a few servers (reduces VTP messages).


-Use VTP passwords to prevent VTP corruption.

VTP Advertisements
-See also VTP advertisements above
-Have revision information (helps determine if it is new information or already known information)
-Never sent to root bridge. Sent downstream to VTP neighbors.

Client generated advertisements:


-Learn about VLANs in network, when they are booted up, and responses from servers.
-Message types in the advertisements:
-Summary, subset, request
-Summary: generated every 300 sec. by default on management VLAN (normally #1), they keep
VLAN information on switches is synchronized.
-Subset: detailed information about specific VLAN.
-Request: generated by clients when they want information about VLANs from a server. The
response is one of the above two message types.
-If no domain is setup, all switches will be in transparent mode.
-Setup first switch as a server, then setup remaining switches as clients, so that they can receive the VTP
information from the server.

Catalyst 5000 VTP Configuration


Switch> (enable) set vtp domain name
Switch> (enable) set vtp domain name mode server | client | transparent
Switch> (enable) set vtp domain name passwd password

Copyright © 2004 cisconotes 47


-Can combine into just one statement
-Domain name up to 32 characters
-Password from 8-64 characters

-2 versions of Cisco VTP: 1 and 2.


-Version 2 switches can revert to version 1 (in a version 1-only switch network).
-Only one version 2 switch need be configured for version 2; Then the remaining switches will change over
to version 2.
-Version 1 is default.
-Version 2 features:
-Supports token ring VLANs
-Forwards and saves unrecognized type lengths.
-Consistency checks when VLANs are adds (verifies number and name of VLAN).
-Transparent mode switches will forward the message even if there is a mismatch in the: version
number, name, password (version 1 switches dropped the message if there was a mismatch).

Switch> (enable) set vtp v2 enable | disable

VTP Pruning
-Reduces unnecessary multicast, broadcast, and unknown destinations across trunk ports: sends that kind of
traffic only to trunk links which have ports belonging to the VLAN in question.
-Disabled by default.
-Requires a management domain.

Switch> (enable) set vtp pruning enable | disable

-All VLANs can be pruned (default), or can specify a range which can be pruned.

Switch> (enable) clear vtp pruneeligible list_of_vlans_or_range


Switch> (enable) set vtp pruneeligible list_of_vlans_or_range

Switch> (enable) show vtp domain

Domain Name Domain Index VTP Version Local Mode Password


---------------- ---------------- ---------------- -------------- ------------
Mydomainname 1 2 server -

Vlan-count Max-vlan-storage Config Revision Notifications


------------- -------------------- -------- ---------- ---------------
11 1023 0 enabled
Last Updated V2 Mode Pruning PruneEligible on Vlans
--------------- ----------- --------- ----------------------------
192.168.1.1 enabled enabled 2-1000

Switch> (enable) show vtp statistics

VTP statistics:
summary advts received 8
subset advts received 6
request advts received 0
summary advts transmitted 501
subset advts transmitted 51
request advts transmitted 28
No of config revision errors 0
No of config digest errors 0

Copyright © 2004 cisconotes 48


VTP pruning statistics:

Join Join Summary advts received from


Trunk Transmitted Received non-pruning-capable device
------ -------------- ----------- ---------------------------------
1/1 341 320 0
1/2 110 115 0

Catalyst 1900/2820/2900XL VTP configuration

SwitchA(config)# vlan database


SwitchA(config-database)# vtp domain domain_name
SwitchA(config-database)# vtp password management_password
SwitchA(config-database)# vtp server | client | transparent
SwitchA(config-database)# [no] vtp pruning
SwitchA(config-database)# [no] vtp v2-mode

show vtp counters --OR--

show vtp status


VTP version :2
Configuration Revision :3
Maximum VLANs supported locally : 66
Number of existing VLANs :6
VTP Operating mode : Server
VTP Domain name : mydomain1
VTP Pruning mode : Disabled
VTP V2 mode : Enabled
VTP Traps Generation : Disabled
MD5 digest : 0xA1 0x90 0x10 0x32 0xCD 0x11
Configuration last modified by 192.168.1.1 at 01-01-00 9:00:01

1900/2820 Switch VTP configuration

SwitchA(config)# vtp [server | transparent | client] [domain domain_name] [password password]


[pruning enable | disable]

show vtp (same as above)


show vtp statistics (same as above)

Available VLAN transport protocols for physical LAN backbones


Fast Ethernet trunked links Æ ISL (Inter-Switch Link) used

ATM trunked links Æ LANE (LAN Emulation)

FDDI trunked links Æ 802.10 VLAN ID field found in FDDI SAID header (also has a protected and clear
header) . VLAN ID found in the 802.10 SAID header can have up to 4.29 billion values

Configuring VLAN Trunk Protocol via menu-based interface


Press V to enter Virtual LAN menu from the main menu then:
N - Enter management domain name (VTP exchanged in a management domain)
[Enter] – display Virtual LAN configuration menu
V – VTP Mode Control menu
S – Server mode

-Each switch now also needs a trunk port configured.

Copyright © 2004 cisconotes 49


-The switch will only learn of other VTP switches via network advertisements.

Spanning Tree (STP) Configuration


Common Spanning Tree (CST) / Mono Spanning tree - one STP instance for the switched network
-IEEE 802.1Q removes CST loops
-Cisco's STP
-Per VLAN Spanning Tree (PVST)
-Shared Spanning Tree (PVST and PVST+)
-PVST used by default on ISL trunks, separate STP instance for each VLAN.
-PVST+ is capable of incorporating CST BPDUs into a PVST

CST
-Only one instance of STP on all VLANs
-STP runs in default management VLAN (usually VLAN 1)
Advantages over PVST:
-Only one set of BPDUs for STP
-Changes tracked for just one instance of STP
Disadvantages
-Likely to make non-optimal paths
-Convergence problems become worse as network grows

PVST
-One instance of STP per VLAN (solves scalability and convergence issues)
-means each VLAN has a root, priorities, port costs, path costs
-STP can be tuned per VLAN.
-STP topology smaller (only uses switches that connect a VLAN together: decreases convergence time and
increase scalability)
-Switch ports with o a VLAN not included in STP topology (but VTP pruning needs to be enabled)
-Any VLAN can be added to a member switch's port easily.

Disadvantages
-Switch multicasts BPDUs on each VLAN.
-Topology database on each switch for each VLAN.
-Optimal configuration requires manual tuning of each VLAN.

PVST+
-Cisco extension to PVST protocol.
-IEEE 802.1Q CST & Cisco's PVST can work on the same switch network.
-Automatic (no configuration).
-Detects CST & PVST, and makes adjustments.

Catalyst 5000 v4.1 software enhancements:


-Tunnel PVST BPDUs across an 802.1Q trunk
-Check for VLAN, port inconsistencies.
-Set port to blocking mode (preventing loops) on inconsistent BPDU receipt.

STP disabling/enabling
-STP enabled by default for all VLANs
-or globally enabled or disabled per VLAN
-STP highly recommended by Cisco even if no loop problems currently exist.

Catalyst 5000
Console> (enable) set spantree enable | disable [vlan_number | all]
sample result: Spantree 10 enabled.

Copyright © 2004 cisconotes 50


Catalyst 2900XL
Switch(config)# [no] spanning-tree [vlan list_of_vlans (separate by spaces) ]

Catalyst 1900/2820
Switch(config)# [no] spanning-tree list_of_vlans

Root Bridge Selection


-Distribution layer switch recommended

Catalyst 5000 STP configuration


Console> (enable) set spantree root [secondary] list_of_vlans [dia diameter_of_net] [hello hello_time]
-Results in priority of root set to: 8192, secondary: 16,384.
-Secondary is a backup for the root bridge
-Dia should be set as in the topic "Convergence", range is 2-7.
-Hello is how often the root bridge sends BPDUs (default: every 2 seconds), range 1-10.

2900XL configuration (access layer device, meaning it's rarely a root bridge)
SwitchA(config)# [no] spanning-tree [vlan list_of_vlans] priority new_priority
-Default switch priority is 32,768 (choose something smaller for the root)

1900/2920
Switch(config)# [no] spantree-template bridge_template_identifier priority new_priority
-Templates are created/modified using this command.
Switch(config)# [no] spantree-template bridge_template_identifier vlan list_of_vlans
-Templates need to be applied using the above command (to take effect).
-Max 4 templates
-Template 1 is the default template used by VLANs. Template 1 does not require the above application
command to take effect.
-If bridge_template_identifier=1, the next command is not necessary.

STP Verification
-This shows a root bridge's configuration ("Bridge ID MAC ADDR" and "Designated Root" are same
proving that it is a root bridge):

Switch_D> (enable) show spantree 1

VLAN 1
Spanning tree enabled
Spanning tree type ieee

Designated Root 00-08-01-ff-fa-fb


Designated Root Priority 32767
Designated Root Cost 30
Designated Root Port 4/1
Root Max Age 20 sec Hello Time 5 sec Forward Delay 15 sec

Bridge ID MAC ADDR 00-08-01-ff-fa-fb


Bridge ID Priority 32767
Bridge Max Age 20 sec Hello Time 2 sec Forward Delay 15 sec

Port Vlan Port-State Cost Priority Fast-Start


1/1 1 forwarding 14 32 disabled
1/2 1 blocking 14 32 disabled
2/1 1 blocking 14 32 disabled
2/2 1 not-connected 14 32 disabled

Copyright © 2004 cisconotes 51


Catalyst 2900XL
SwitchA# show spanning-tree [vlan vlan_number]
-it will show vlan 1 if vlan_number is not specified, like the below sample output:

Spanning tree 1 is executing the IEEE compatible Spanning Tree protocol


Bridge Identifier has priority 32768, address 0010.1add.0011
Configure hello time 2, max age 20, forward delay 15
Current root has priority 32768, address 0010.1ade.4400
Root port is 10, cost of root path is 10
Topology change flag not set, detected flag not set, changes 1
Times: hold 1, topology change 12, notification 4
hello 2, max age 20, forward delay 15
Timers: hello 0, topology change 0, notification 0

Interface Fa0/1 in Spanning tree 1 is down


Port path cost 100, Port priority 128
Designated root has priority 32768, 0010.1ade.4400
Designated bridge has priority 32768, address 0010.1add.0011
Designated port is 1, path cost 10
Timers: message age 0, forward delay 0, hold 0
BPDU: sent 0, received 0

-for Catalyst 1900/2820 use show spantree (Catalyst 5000's command). Output is same format as above.

Wide Area Networking


-WAN protocols are diverse. Similar to LAN technology, in that they use layer 1 and 2 to encapsulate and
move data from one device to the next device of the same device type.

Serial transmission - transmission one bit at a time through a single wire. Reception is done on a separate
single wire, one bit at a time. Virtually all WANs use serial ports that use serial transmission.
Parallel transmission - 8 bits at a time through 8 wires. Alternatively, this is done in multiples of 8.

Synchronous transmission - uses a clocking signal to define beginning & end of bits (or frames).
Asynchronous transmission - does not use a clocking signal to define beginning & end of bits (or frames).

Leased lines - point to point dedicated connections. Most are synchronous serial connections up to
45Mbps. No setup/tear-down procedure needed for communication.

Circuit switching - setup like a phone call (choosing physical circuits) needed for communication,
Dial-up modems and ISDN use this.

Packet switching - Packets from many different companies are switched by a packet switch provider.
Shares bandwidth with other companies to save money. Can only be used if traffic is bursty. If traffic is
constant, you need a leased line instead. Frame relay, and X.25 are two standards that are packet switched
by providers.

Plain Old Telephone Service (POTS)


Most commonly used WAN technology, and oldest and largest. More phone than internet users in world.

Demarc – boundary between customer’s wiring and service provider’s wiring. It’s demarcation point, or
end of responsibility for the service provider.

CPE – Customer premise equipment – all wiring and equipment on customer’s side of demarc.

Copyright © 2004 cisconotes 52


Local loop – The wiring from the demarc to the CO. Sometimes called last mile.

CO (Central Office) – point where local loop gains access to the service providers high-speed trunk lines.
AKA POP – point of presence.

Toll Network – Trunk lines in a WAN provider. Many switches are connected to them.

Physical media
-T1s are almost always STP cabling with RJ45 connectors on each end.
-T3s are almost always coax cabling with BNC connectors on each end.
-Frame relay uses various cabling, typically 2 pair UTP, with ”smartjacks” (which are RJs other than RJ11
and RJ45)

Synchronous Data-Link Control (SDLC)


-IBM originally developed it for use in SNA protocol – SDLC is the chief WAN link layer protocol for
SNA nets
-Predecessor to HDLC, & the basis for HDLC, LAPB, and others.
-Bit-synchronous Data Link layer protocol (expanded efficiency, flexibility, sometimes greater speed.
-Cisco routers can be configured with SDLC across serial interfaces to connect to remote building or office
with SNA protocol suite.
-Supports many topologies:
-Point to point and multipoint links
-Bounded and unbounded media
-Half-duplex and full-duplex transmission
-Circuit and packet switched networks

-Supports two network node types:


primary station – control operation of other stations, poll secondaries in predetermined order. Setup, tear
down, and manager links.

secondary station – controlled by primary station. If a secondary is polled, it is allowed to transmit only to
the primary station (“speak when spoken to”- IBM).

High-Level Data-Link Control (HDLC)


-Popular ISO standard; ISO modified SDLC to create HDLC
-Most commonly used encapsulation on serial interfaces which are in turn connected to T1s, T3s, etc.
-PPP is the 2nd most common encapsulation found on T1s, T3s, etc.. See example at the very end of the
document.
-Bit-oriented link-layer protocol that specifies an encapsulation method for data on synchronous serial data
links.
-ITU-T tweaked HDLC to create LAP then LAPB Link Access Procedure Balanced
-IEEE worked on HDLC and created 802.2 from it.
-Default encapsulation used by Cisco routers over synchronous serial links.
-Cisco’s HDLC is proprietary, as is every other vendors.

Show int s0
Serial0 is up, line protocol is up
Hardware is HD64570
Internet address is 172.16.20.1/24
MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec, rely 255/255, load 1/255
Encapsulation HDLC, loopback not set, keepalive set (10 sec)
Last input 00:00:05, output 00:00:01, output hang never
Last clearing of “show interface” counters never
[Queues, then conversations, packet statistics, carrier transitions, and DSR, DTR, RTS, CTS]

Copyright © 2004 cisconotes 53


Transfer Modes Supported by HDLC
-HDLC uses transfer modes to define media access on the WAN.

Normal response mode (NRM)


-Implemented in SDLC
-Secondaries can’t communicate with primary until primary asks.

Asynchronous response mode (ARM)


-Secondaries can communicate with primary without permission.

Asynchronous balance mode (ABM)


-Combined node (act as primary or secondary). All comm. takes place between combined nodes.
-Combined stations can originate transmissions without permission.

Link Access Procedure Balance (LAPB)


-Integrated into X.25 protocol stack.
-Same frame format, frame types, and field functions as SDLC & HDLC.
-Confined to ABM transfer mode with which you can establish circuits without DTE or DCE.
-Devices that initiate communication (primaries). Those that respond (secondaries).

Dial-on-Demand Routing (DDR)


-Allows WAN links to be used selectively. Admin defines interesting traffic (this initiates WAN link).
Interesting traffic is defined by access lists – a lot of flexibility; great for minimizing traffic for expensive
per-min IDSN WANs.

-DDR great for per-min ISP links, and fully functional backup system.
-Versatile DDR can be used over several different types of connections (IOS v9+): IP, IPX, AppleTalk,
DECnet, OSI/CLNS, and others.
-DDR can be used over different interfaces: sync ser., asynchronous serial, ISDN.
-No dynamic routing information learned from these interfaces (they are usu. inactive) -> Need static routes

-Static routes have admin. distance of 0 (override any dynamically learned route)! – acts like a primary
route! Use: ip route 10.1.2.3 255.255.255.0 10.5.6.7 <1-255> to specify admin distance. Choose 200+
because EIGP default is 170; then DDR will only be used if all other routes absent or unreachable

-Bad idea to use OSPF, BGP, or IS-IS in DDR environment. These protocols require ACK from neighbor
before updates are sent - DDR is usually inactive when ACK needs to be sent. USE DDR as backup
connection with static routes

-DDR follows 4 steps to establish a link:


(routers on both sides must have DDR on interfaces – duh!)
1 – router verifies a route exists to destination
2 – via the static route, router finds the DDR interface connected to destination.
3 – router checks to see if DDR is already active and connected to destination
4 – router determines if packet is interesting or uninteresting based on access list applied to int.
If packet is interesting, and not already connected, will make connection, then send packets.

X.25
-Born before today’s digital networks – built for noisy analog circuits, it’s overbuilt for today’s needs.
-Addressing defined by X.121 (1-14 decimal digits). 1st 4 bits – DNIC Data Network Identifier Code,
remaining bits set by administrator.

-Defines point-to-point communications between DTEs and DCEs.


DTE – router, terminal, PCs, bridges

Copyright © 2004 cisconotes 54


DCE – modem or CSU/DSU. DCE’s are used as relays between DTE’s.
Or even packet switches: almost all are carrier-owned.

-DCE connects to the service provider’s network with the goal of establishing a virtual circuit between two
DTE devices.
-X.25 supports both switched and permanent virtual circuits.
-X.25 heavily used in PSN packet-switched networks of telcos that charge on use.
In 70’s need for WAN protocols that could connect PDN public data networks, X.25 now administered by
ITU-T.

3 Categories of X.25 Network Devices:


Data Terminal Equipment (DTE): End systems that communicate over an X.25 network (host systems,
terminal, and PCs that belong to the individual subscriber) and are present at the same site

Data Circuit-Terminating Equipment (DCE): Specific communications equipment such as packet


switches that interface between a packet switching exchange (PSE) and DTE devices. Typically these are
only in carrier facilities.

Packet Switching Exchange (PSE): Switches that constitute the majority of a carrier’s network and
handle the transfer of data between DTE devices via the X.25 packet-switched network.

X.25 Sessions
Sessions used for DTE – DTE communication:

1. DTE device contacts another to request a communication session


2. Receiving DTE device accepts or refuses
3. If accepted, both systems begin full-duplex information transfer
4. Either DTE device can terminate the connection

After termination, further comm. requires making a new session.

Virtual Circuits Over an X.25 Network


-Virtual circuits are logical connections (not physical).
-They are logical, bi-directional paths from one DTE to another over a X.25 network.
-Connection can pass through many PSEs and DCE devices.
-A bunch of virtual circuits can be multiplexed onto one physical circuit, then later de-multiplexed.
-X.25 uses two types of virtual circuits: SVC and PVC

SVC - Switched Virtual Circuit


-Temporary connection for intermittent data transfers: 2 DTEs establish, maintain, terminate a session
every time.

PVC - Permanent Virtual Circuit


-Established and used to recurrent, steady data transfer. DTE can transmit whenever necessary (session
always ready).

X.25 Protocol Suite


-Maps to first 3 layers.
-Packet Layer Protocol (PLP)
-Link Access Procedure Balanced (LAPB)
-X.21bis and other physical layer serial interfaces (EIA/TIA 232, EIA/TIA-449, EIA-530, G.703)

Packet Layer Protocol (PLP)


-X.25’s network layer protocol
-Maintains packet exchanges between DTE devices across virtual circuits
-Can also run over LLC2 implementations on LANs

Copyright © 2004 cisconotes 55


-Can also run over ISDN interfaces running LAP on D channel (LAPD).

Five modes of operation


Call Setup Mode
-For setup of SVCs between DTEs. PLP initially uses X.121’s addressing for setup.
-Different virtual circuits can be in different modes at the same time because call setup mode is
-Deployed as individual virtual circuits require.
-SVC only.
Data Transfer Mode
-For data transfer between 2 DTE devices via virtual circuit.
-Segmentation, reassembly, bit padding, and error and flow control here.
-On a per-virtual-circuit basis.
-SVC & PVC.
Idle Mode
-When virtual circuit is established, but no transfer of data occurring.
-On a per-virtual-circuit basis.
-SVC only.
Call Clearing Mode
-Terminate communication sessions between DTEs and SVCs
-On a per-virtual-circuit basis.
-SVC only.
Restarting Mode
-Synchronizes transmission between DCE device (locally connected) and a DTE device.
-Communication and packet framing between DTE and DCE devices happen here.
-Affects ALL DTE device’s established virtual circuits (NOT per-virtual-circuit basis)

Link Access Procedure Balanced (LAPB)


-Make sure frames are error free and properly sequenced.
-Bit-oriented protocol. (full duplex, more efficient, and reliable, and doesn’t have to be byte-size).

3 different frame types

Information Frames (I-frames)


-Transmit upper layer information & control information.
-Send and receive sequence numbers.
-Sequencing, flow control, error detection, and recovery.

Supervisory Frames (S-frames)


-Has control information.
-Carries sequence numbers
-Requests for and suspension of transmission
-Report on status and ACKs I-frame receipt.

Unnumbered Frames (U-Frames)


-Has control information.
-Link setup and disconnection.
-Error reporting
-No sequence numbers.

X.21bis
-Used in X.25 at physical layer –electrical and mechanical properties for media.
-Activation / Deactivation of physical media between DTE, DCEs.
-Max 19.2Kbps
-Point-to-point, synchronous, full-duplex, four-wire media.

Copyright © 2004 cisconotes 56


X.25 on Cisco Routers

encap x25 while in interface configuration mode.


x25 ? –configuration parameters
Accept-reverse Accept all reverse charged calls
Address Set interface X.121 address
Alias Define an alias address pattern
Default Set protocol for calls with unknown Call User Data
Facility Set explicit facilities for originated calls
Hic Set highest incoming channel
Hoc Set highest outgoing channel
Hold-queue Set limit on packets queued per circuit
Hold-vc-timer Set time to prevent calls to a failed destination.

-X.121 addresses aren’t burned into ROM, need to assign it to an X.25 serial int.
int s0
x25 address 12345678

-Default packet size 128 doesn’t work with every vendor.


-Correct Input Packet Size IPS x25 ips
-Correct Output Packet Size OPS x25 ops

-Some packets are used by flow control mechanisms.


Default window size is 2
x25 win 7 – x25 window input size
x25 wout 7 –x25 window output size

sh int s0
Serial 0 is admin….
..
MTU…
Encapsulation X25, loopback not set
LAPB DTE, modulo 8, k 7, N1 12056, N2 20
T1 3000, interface outage (partial T3) 0, T4 0
State DISCONNECT, VS 0, VR 0, Remote VR 0,
Retransmissions 0
Queues: U/S frames 0, I frames 0, unack. 0, reTx 0
IFRAMEs 0/0 RNRs 0/0 REJs 0/0 SABM/Es 0/0 FRMRs 0/0 DISCs 0/0
X25 DTE, address 12345678, state R1, modulo 8,k timer 0
Defaults: cisco encapsulation, idle 0, nvc 1
Input/output window sizes 7/7, packet sized 256/256
Timers: T20 180, T21 200, T22 180, T23 180, TH 0
Channels: Incoming-only none, Two-way 1-1024, Outgoing-only none
RESTARTs 0/0 CALLs 0+0/0+0/0+0 DIAGs 0/0

Frame Relay
-One of the most popular WAN protocols. Inexpensive way to connect large number of sites.
-A reason for popularity shared bandwidth through frame relay cloud.
-Data traffic is bursty, so 100’s of users can share various leased lines, etc.
-Less expensive for telcos and customers this way.
-Two sites can be connected with up to T1 speeds via frame relay cloud; will be T1 as long as not saturated.
-Much cheaper than T1 in this case.
-The frame relay provider would allow bursting to T1 in this scenario,
-And also provide a CIR (committed information rate) – a minimum guarantee.
-1.5 CIR would loose price advantage over a 1.5 leased line.

Copyright © 2004 cisconotes 57


-Appealing price: Pay for what you need, but can burst much more

Frame Relay Features


-High-performance WAN protocol operating on physical and data-link layer.
-Originally used on ISDN interfaces, now on a variety of interfaces.
-Gives a comm. interface between DTE and DCE.
-Popular opinion that frame relay faster, more efficient than X.25 because error checking done by higher-
layer protocols..
-Connection-oriented data link layer communication via virtual circuits (like X.25).
-A frame relay virtual circuit is a logical connection between two DTEs across a packet-switched network
(PSN), and is identified by a DLCI.
-Uses both PVCs and SVCs, but most are PVCs.

Frame Relay with Cisco Routers


int s0
encapsulation frame-relay –two types: cisco (default) and ietf

-Use cisco if connecting between 2 Cisco devices, otherwise use ietf.

DLCIs (Data-Link Connection Identifiers)


-Frame relay virtual circuits are identified by DLCIs.
-Frame relay svc provider assigned DLCIs
-Frame relay can have multipoint interface (multiple DLCIs)
-Must map IP address to DLCI before IP works.
-Frame relay uses DLCIs the same way that X.25 uses X.121 addresses.
-DLCI has either global or local (usu.) meaning in frame-relay network.
-Local DLCI has different DLCI value when referring to same connection.

RouterA(config-if)#Frame-relay interface-dlci 16 <16-1007>

Local Management Interface (LMI)


-Developed 1990 Cisco, StrataCom, Northern Telecom, and DEC “Gang-of-Four LMI” or “Cisco LMI”
-They added extensions to basic frame-relay protocol from CCIT.
-Allows network Devices to communicate more easily with frame relay network.

LMI messages provide:


-Current DLCI values
-Global or Local significance
-Status of virtual circuits

-Default LMI type is Cisco. ANSI, Q.933a (ITU-T) also avail, check with frame relay provider.
frame-relay [lmi-type]
lmi-types available: cisco, ansi, Q933a

All standard signaling formats are supported by above types.

LMI TYPE AUTODETECTED IN IOS v11.2

Subinterfaces
-Can have each virtual circuit on a single interface be treated as a single interface – subinterface.
-Subinterface is a virtual interface defined by IOS software.

-Each subinterface can have different network-layer characteristics.(and therefore virtual circuit, also), i.e.
IP on one, and IPX on another.

Copyright © 2004 cisconotes 58


int s0.16 [multipoint] [point-to-point]

-Chose 16 to match DLCI number.


-Multipoint means supports multiple virtual circuits

Mapping Frame Relay 2 ways:

1.
encap frame-relay ietf
no inverse-arp
ip address…
frame-relay map ip 172.16.30.17 30 cisco broadcast
frame-relay map ip 172.16.30.18 50 broadcast – will use ietf from above
frame-relay map ip 172.16.30.19 40 –will use ietf from above

-Map command only way to mix cisco and ietf


-Broadcast – forward broadcasts for this interface to the virtual circuit.

2.
Use inverse-arp function – dynamic mapping of IP address to DLCI number
int s0.16
encap frame-relay ietf
ip address 172.16.30.1 255.255.255.0

Note: this method is not as stable as map command! – configuration errors occur because virtual circuits
can be mapped to unknown devices.

Monitoring Frame Relay


sho frame [ip | lmi | map | pvc | route | traffic]
sho frame pvc –shows configured PVCs and DLCIs, status, and packet statistics.

sho int s0 also shows general (fewer LMI stats) packet statistics after “Encapsulation FRAME-RELAY”.

Point-to-Point Protocol (PPP)


-Data-link layer used over asynchronous dial-up or ISDN.
-LCP (Link Control Protocol) to build and maintaining Data Link connections:
-authentication using either PAP Password Authentication Protocol or CHAP Challenge-
Handshake Authentication Protocol.
-Data compression
-Supports IP, IPX, AppleTalk, DECnet, OSI/CLNS and Transparent bridging via NCP (Network
Control Protocol) – a family of protocols, one for each layer 3 protocols supported.

-Those features not available in SLIP (predecessor).

Configuration
In interface configuration mode:
Encapsulation ppp

-Additional commands:
ppp [authentication] | [bridge] | [chap] | [compression] | [multilink] | [pap] | [quality] | [reliable-link] |
[timeout] | [use-tacacs]

show int s0

Hardware is HD64570

Copyright © 2004 cisconotes 59



Encapsulation PPP, loopback not set, keepalive set (10 sec)
LCP Open – to build and maintain connections.
Listen: IPXCP
Open: IPCP, CDPCPO, ATCP – IP, CDP, AppleTalk support from NCP.

ISDN (Integrated Services Digital Network)
-Digital service designed to run over existing phone networks.
-Typical ISDN apps: high-speed image, high speed file trans., videoconferencing, multiple links for
telecommuters

ISDN - set of protocols from telcos. Simultaneous data, text, voice, music, graphics, video over existing
telephone. ITU-T standardized to cover physical, data link, & network layers.

ISDN Terminals
Devices connecting to ISDN network are terminals:

TE1 – Terminal equipment type 1 understands ISDN standards.


TE2 – Terminal equipment type 2 predate ISDN standards – must use terminal adapter TA to connect to an
NT1.
NT1 – ISDN physical layer spec. ends here and allows end-user devices to connect to it.
NT2 – ISDN physical layer and devices which includes the provider's equipment. Rarely found at
customer premises.

ISDN Reference Points


4 of them to define logical interfaces
R reference point – point between TE2 and TA.
S reference point – point between NT1 and either a TA or a TE1.
T reference point – point between NT1 and NT2 (which is also found in exactly the same place as the S
reference point.)
U reference point – point between NT1 devices and line-termination equipment inside a carrier network.
(only in North A. where NT1 function isn’t provided by carrier network)

ITU-T Defined Protocols for Diverse Issues


Start with letter:
E ISDN on Existing telephone network.
I Info: concepts, terminology, services.
Q Switching and signaling.

Basic Rate Interface


2B channels + 1D channel

B-channel is 64Kbps data


D-channel is 16Kbps control & signaling information (physical, data link, network)
Also has 48K framing control (up to total grand total of 192Kbps).

Need a SPID to configure each B-channel (just like a phone number for each B channel).
ISDN devices send SPID to ISDN switch in order to gain PRI or BRI access.
Sometimes SPIDs are auto-sensed (used a special “config SPID” command)

Without SPID, many ISDN switches don’t allow an ISDN device to place a call on network.

Primary Rate Interface


Aka 23B+ 1D Æup to 1.544Mbps (= to T1 speed) in North A., Japan
30B+ 1 64K D Æ up to 2.048Mbps (= to E1 speed). in Europe, Australia, rest of world.

Copyright © 2004 cisconotes 60


ISDN with Cisco Routers
-Must purchase NT1 or an ISDN modem & connect to serial interface, unless router has BRI interface.
-Router without BRI int. is a TE1, one that req. a TA is called a TE2.

-ISDN supports almost all upper-layer protocols: IP, IPX, AppleTalk, etc.
-Choose PPP, HDLC, or LAPD as encapsulation protocol.

-Know switch type when configuring, use isdn switch-type ? to see which are avail. Each manufacturer
has proprietary protocol. for signaling.

-For each ISDN BRI, specify SPIDs:


config t
isdn switch-type basic-dms100
int bri0
encap ppp – must be ppp: ISDN uses it for digital phone call
isdn spid 775456721
isdn spid 775456722

-For each PRI interface, specify data-link-specific information for T1 controller. This is how PRI
communicates with ISDN switch.

RouterA(Config)#controller T1 1/0
RouterA(Config-if)#framing esf
RouterA(Config-if)#linecode b8zs
RouterA(Config-if)#pri-group timeslots 1-24

-Extended SuperFrame Framing (ESF) is used on T1 circuits: 24 frames, 192 data bits each, #193 for
timing, other.

-B8ZS binary eight-zero substitution line-coding that guarantees density of 1s over a link by substituting a
special code whenever eight consecutive 0s are sent, then removing code at the remote end.

-Timeslots 1-24: defines ISDN PRI timeslots (1-24)

Sample Show Configuration Command on a T1


-Example of show interface command on 3 different actively used T1s on the same router:

RouterA# show int s0/0


Serial0/0 is up, line protocol is up
Hardware is PQUICC with Fractional T1 CSU/DSU
MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec,
reliability 255/255, txload 51/255, rxload 159/255
Encapsulation PPP, loopback not set
Keepalive set (10 sec)
LCP Open, multilink Open
Last input 00:00:00, output 00:00:00, output hang never
Last clearing of "show interface" counters 00:00:23
Queueing strategy: fifo
Output queue 0/40, 0 drops; input queue 0/75, 0 drops
5 minute input rate 964000 bits/sec, 127 packets/sec
5 minute output rate 311000 bits/sec, 104 packets/sec
3065 packets input, 2894747 bytes, 0 no buffer
Received 0 broadcasts, 0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort

Copyright © 2004 cisconotes 61


2555 packets output, 1163617 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 output buffer failures, 0 output buffers swapped out
0 carrier transitions
DCD=up DSR=up DTR=up RTS=up CTS=up

RouterA# show int s0/1


Serial0/1 is up, line protocol is up
Hardware is PQUICC with Fractional T1 CSU/DSU
MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec,
reliability 255/255, txload 50/255, rxload 159/255
Encapsulation PPP, loopback not set
Keepalive set (10 sec)
LCP Open, multilink Open
Last input 00:00:00, output 00:00:00, output hang never
Last clearing of "show interface" counters 00:00:24
Queueing strategy: fifo
Output queue 0/40, 0 drops; input queue 0/75, 0 drops
5 minute input rate 965000 bits/sec, 127 packets/sec
5 minute output rate 307000 bits/sec, 104 packets/sec
3180 packets input, 2980276 bytes, 0 no buffer
Received 0 broadcasts, 0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort
2651 packets output, 1197081 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 output buffer failures, 0 output buffers swapped out
0 carrier transitions
DCD=up DSR=up DTR=up RTS=up CTS=up

RouterA# show int s1/0


Serial1/0 is up, line protocol is up
Hardware is DSCC4 with integrated T1 CSU/DSU
MTU 1500 bytes, BW 1536 Kbit, DLY 20000 usec,
reliability 255/255, txload 51/255, rxload 160/255
Encapsulation PPP, loopback not set
Keepalive set (10 sec)
LCP Open, multilink Open
Last input 00:00:00, output 00:00:00, output hang never
Last clearing of "show interface" counters 00:00:25
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: weighted fair [suspended, using FIFO]
FIFO output queue 0/40, 0 drops
5 minute input rate 964000 bits/sec, 127 packets/sec
5 minute output rate 309000 bits/sec, 104 packets/sec
3315 packets input, 3127346 bytes, 0 no buffer
Received 0 broadcasts, 0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort
2768 packets output, 1283898 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 output buffer failures, 0 output buffers swapped out
0 carrier transitions
DCD=up DSR=up DTR=up RTS=up CTS=up

Copyright © 2004 cisconotes 62