Using Frame Relay   13

  Frame Relay is a method of encapsulating network information that allows for fast delivery and high line utilization. PortMaster routers support Frame Relay over synchronous ports.

  This chapter uses an example to demonstrate how to configure the PortMaster 4 to connect to a synchronous line using Frame Relay. This chapter also explains how to configure Frame Relay subinterfaces.

  The following topics are discussed:

•  "Overview of Frame Relay" on page 13-1
•  "Frame Relay Configuration on the PortMaster" on page 13-3
•  "Configuring a T1 or E1 Line for Frame Relay" on page 13-4
•  "Troubleshooting a Frame Relay Configuration" on page 13-7
•  "Frame Relay Subinterfaces" on page 13-8

  See the PortMaster 4 Command Line Reference for more detailed command descriptions and instructions.

  You can also configure the PortMaster 4 using the PMVision application for Microsoft Windows, UNIX, and other platforms supporting the Java Virtual Machine (JVM). PMVision replaces the PMconsole interface to ComOS.

       Overview of Frame Relay

  Frame Relay is a switched digital service that supports multiple virtual circuits, simultaneously connected to a site by a single physical circuit. Each site requires only one physical circuit into the Frame Relay network--usually referred to as a cloud--but can have several virtual circuits to reach other sites attached to the cloud.

  You configure synchronous ports on the PortMaster 4 to support Frame Relay connections. As opposed to a dedicated or leased line, a Frame Relay connection can be thought of as a virtual switch.

       PVCs and DLCIs

  Permanent virtual circuits (PVCs) are used to form a connection between any two devices attached to a Frame Relay cloud. Each PVC is given a unique number on each physical circuit along the path between the two devices. This unique number is called a data link connection identifier (DLCI). The DLCI is automatically changed to the PVC number of the next physical circuit as it passes through each switch along the path. A DLCI is different from a network address in that it identifies a circuit in both directions, not a particular endpoint. A frame contains only one DLCI, not a source and destination.

  In general, the only DLCI numbers you see are those numbers assigned to the physical circuits on the perimeter of the Frame Relay cloud.

       Line Speed

  The physical circuit between point A and the network must be ordered with a certain line speed. This speed is the maximum physical bandwidth for your connection to the Frame Relay network. Expansion beyond this limit is not possible without a hardware change and a new circuit installation.

       Port Speed

  The connection into the telecommunications provider's Frame Relay network must be ordered at a particular port speed, which is the maximum bandwidth rate that the telecommunications provider accepts from your connection. This number must be less than or equal to the line speed. This speed is the maximum rate at which you can transmit data to any of your PVCs under any circumstances. The port speed differs from line speed only in that it can be upgraded through software without a circuit installation or hardware change.

       CIR and Burst Speed

  Each PVC has a property known as committed information rate (CIR), which represents the guaranteed minimum bandwidth available to the particular PVC under all conditions. In some implementations, an additional property can be assigned to a PVC, known as burst speed or maximum burst. This speed represents the highest rate at which data is allowed to flow over a given PVC, regardless of bandwidth availability.

       Discarding Frames

  The PortMaster 4 transmits as much data on the serial port as it can for any PVC that has traffic, regardless of CIR. The Frame Relay switch passes on as much of the data as possible to the next link. However, once a particular PVC has transmitted its CIR-worth of bits each second, the switch marks any additional frames as "discard eligible." If the switch receives more frames than it can pass along, the frames are automatically discarded in the following order:

•  Frames that would be marked discard eligible even if they are forwarded
•  Frames received that were marked as discard eligible

  If the switch must discard other frames, the behavior is undefined. In this case, the Frame Relay network is improperly configured because the CIR total exceeds the line speed or port speed.

       Ordering Frame Relay Service

  In general, when ordering Frame Relay service for a private network, order large-bandwidth physical circuits (for example, T1) with a port speed appropriate to your application, and a CIR that is high enough to provide minimally acceptable performance for your application. In most cases, ordering according to these criteria provides service that is close to your port speed. The CIR is a guaranteed minimum throughput, not a maximum limit. Port speed is the maximum limit.

       LMI Types

  The following Frame Relay terms relate to network management. The Frame Relay specification supports automatic network status updates, which are exchanged between adjacent devices in the Frame Relay network. These status updates are known as the Local Management Interface (LMI). Two forms of LMI are available in the PortMaster: Cisco LMI, which is commonly referred to as LMI, and ANSI T1.617 Annex D LMI, which is commonly referred to as Annex-D.

  Generally, your telecommunications provider offers three LMI options for your physical circuit: LMI, Annex-D, or none. Because LMI exists only between your router and the switch to which your physical circuit connects, it does not need to match what the remote ends of your PVCs are using. However, your circuit LMI must match the configuration on your PortMaster. Generally, Annex-D is recommended because it is a more feature-rich and robust version of LMI.

       Frame Relay Configuration on the PortMaster

  You configure Frame Relay by selecting the Frame Relay protocol, setting the IP address of the port, and specifying the DLCIs during the synchronous port configuration.

  Alternatively, the PortMaster can discover DLCIs dynamically with LMI or Annex-D and learn the IP addresses of the other routers through Inverse ARP if the other routers on your Frame Relay cloud support Inverse ARP as specified in RFC 1490. In this configuration, the PortMaster sends an LMI status request every 10 configurable seconds by default. Every sixth request is a full status request, and the others are keepalives. In this configuration, the port state is CONNECTING until it receives three replies from the switch; then the port state becomes ESTABLISHED. After six unanswered requests, the PortMaster resets the port.

  Note ¯ All synchronous ports require an external clock signal-either from the device to which the PortMaster is connected or from the telephone company-to regulate the port speed.

       Enabling LMI

  You can specify whether the PortMaster accepts Local Management Interface (LMI) frames from the attached Frame Relay switch. If LMI is enabled on the switch, you must enable LMI on the PortMaster. The default keepalive value is 10 seconds. However, if your telephone company chooses another keepalive value, change this value as they instruct you. Enabling LMI causes the DLCI list to be completed automatically. If the attached switch uses an interval keepalive timer different from the Frame Relay default, be sure the keepalive timer on the PortMaster matches that of the attached switch.

  To enable LMI, use the following command:

  Command> set  W1 lmi  Seconds

  Note ¯ Contact your Frame Relay carrier to determine which keepalive timer they use, LMI or Annex-D.

  To enable LMI, use the following command:

  Command> set  W1 lmi  Seconds

       Enabling Annex-D

  The PortMaster also accepts the Annex-D polling interval. The Annex-D default value is 10 seconds. However, if your telephone company chooses another keepalive value, change this value as they instruct you. Enabling LMI causes the DLCI list to be completed automatically. Setting the keepalive value to 0 (zero) seconds, or enabling LMI, disables Annex-D.

  Note ¯ Contact your Frame Relay carrier to determine which keepalive they are using, LMI or Annex-D.

  To enable Annex-D, use the following command:

  Command> set  W1 annex-d  Seconds

       Listing DLCIs for Frame Relay Access

  If LMI or Annex-D is not used, you must enter the DLCI list manually. The DLCI list is a list of DLCIs that are accessible through the Frame Relay network by this interface. The PortMaster uses Inverse ARP to learn the IP addresses of routers attached to the PVCs represented by the specified DLCIs, if those routers support Inverse ARP. Alternatively, you can specify single IP addresses or IPX network numbers by appending a colon (:) and address or number after the DLCI. See the PortMaster 4 Command Line Reference for more information.

  The DLCI list can be provided by your Frame Relay carrier. For dynamically learned lists, 32 PVCs are allowed. Only 16 PVCs can be specified if the DLCI and IP address are entered. If you specify only DLCIs, you can list 24. When the PVC and IP address are specified, the PortMaster statically configures these entries into its ARP table.

  To enter the DLCI list manually, use the following command:

  Command> add dlci |ipxdlci  W1 Dlci :[Ipaddress|Ipxnode]

  For information on Frame Relay subinterfaces see "Frame Relay Subinterfaces" on page 13-8.

       Configuring a T1 or E1 Line for Frame Relay

  You configure Frame Relay on the PortMaster 4 by configuring a T1 line on the Quad T1 board or an E1 line on the Tri E1 board.

  This section describes how to configure one end of a Frame Relay connection. Because configuration for both ends of the connection is the same, you can use the procedure in this section as a guide for configuring both ends of your Frame Relay connection.

  Before you configure a line for Frame Relay, you must configure global settings on the PortMaster 4, set the gateway routers, and configure the Ethernet interface. Follow the procedures in this section to configure a T1 or E1 line for Frame Relay service.

       Configuring Global and Ethernet Settings

  You configure global and Ethernet settings on the PortMaster 4 from the manager module (slot 4), which is the default view. See Chapter 2, "Configuring Global Settings," and Chapter 4, "Configuring an Ethernet Interface," for more information.

  Note ¯ Lucent recommends that you configure Ether1 if you configure only one Ethernet interface. If you configure both, you must connect them to separate Ethernet segments. Ether0 operates at 10Mbps and is physically on the manager board. Use Ether1 for netboots, SNMP, RADIUS, and syslog .

  Follow this procedure to configure the global and Ethernet settings on the PortMaster 4.

  1. Set the IP address of the gateway router.

  Command> set gateway Ipaddress

  2. Save the configuration to nonvolatile RAM.

  Command> save  all 

  3. Set the IP address of the Ethernet interface.

  Command> set ether1 address Ipaddress

  4. Set the netmask of the Ethernet interface.

  Command> set ether1 netmask Netmask

  5. Set broadcast to high.

  Command> set ether1 broadcast high

  6. Enable RIP routing on the interface.

  Command> set ether1 rip on

  7. Save the configuration to nonvolatile RAM.

  Command> save all

  8. Reset the slot

  Because the Ethernet board on the manager module is in logical slot 10, you activate the Ether1 configuration by resetting slot 10.

  Command> reset slot10

       Configuring the Synchronous WAN Port

  Follow this procedure to configure the synchronous WAN port on the PortMaster 4:

  1. Enter the show boards command to determine the identification number of the line board you want to configure.

  The board identification number is the same as the number of the slot in which the T1 or E1 module is installed.

  Command> show boards

  2. Set the view to a slot with an installed T1 or E1 line board.

  Setting the view for a specific board gives you administrative access to that board.

  Command> set view Slotnumber

  Command Slotnumber>

  The slot number appears in the prompt to identify the board you are configuring.

  3. Set the network type.

  Command Slotnumber> set W1 network hardwired

  4. Set the protocol.

  Command Slotnumber>  set W1 protocol frame

  5. Set the IP address of the port.

  Command Slotnumber> set W1 address Ipaddress

  6. Set the netmask.

  Command Slotnumber> set W1 netmask Netmask

  7. Monitor the carrier detect signal.

  Command Slotnumber> set W1 cd  on 

  8. Enable RIP on the interface.

  Command Slotnumber> set W1 rip broadcast 

  9. Set the Annex-D polling interval.

  You can use LMI instead of Annex-D.

  Command Slotnumber> set W1 annex-d Seconds

  10. Add a data link connection identifier (DLCI).

  You do not need to set a DLCI list if the remote router supports Inverse ARP.

  Command Slotnumber> add dlci W1 Dlci :Ipaddress

  11. Reset the slot.

  Command Slotnumber> reset W1

  12. Save the configuration.

  Command Slotnumber> save all

  If LMI or Annex-D is set, the PortMaster receives DLCI information in the full status update messages from the Frame Relay switch. The PortMaster then attempts to discover IP addresses of other routers using Inverse ARP. You can set DLCI lists statically as well. The show arp frm1  command lists both the static and dynamic DLCI lists for the W1 port.

  If Annex-D is available from your carrier for a new connection, it is preferable to LMI.

  To connect to Cisco routers using Frame Relay, the Cisco router must be set to use encapsulation  frame-relay ietf  for the serial interface; otherwise, the Cisco frame-relay map  command for your DLCI must have the ietf  keyword appended.

  For more information about synchronous ports, refer to Chapter 6, "Configuring a Synchronous WAN Port."

       Troubleshooting a Frame Relay Configuration

  Most synchronous configurations establish links with very little trouble if you have configured the PortMaster using information from your carrier. If you are having problems, use the information in this section to debug your configuration.

  If you are having trouble with a Frame Relay connection, do the following:

•  Wait a few moments. The process of establishing a Frame Relay link, learning the DLCI list, and learning the IP address through Inverse ARP can sometimes take a few moments.
•  Verify that the error counters are 0 except for abort errors. If your counters are nonzero, the problem is external to the PortMaster.
•  Verify that you are using the correct cables and that they are attached securely to the correct port. Not all WAN ports are capable of the same speeds.
•  Enter the following two commands to view the LMI or Annex-D keepalives:

    Command> set console  C0
  Command> set debug 0x51 

    After you verify that the proper keepalives are being received, enter the following commands to turn off the debug utility:

    Command> set debug off
   Command> reset console 

•  If you have a Cisco router on the other end of your connection, verify that it is set for encapsulation  frame-relay ietf  for the serial interface; otherwise, the Cisco frame-relay map  command for your DLCI must have the ietf  keyword appended.

       Frame Relay Subinterfaces

  PortMaster routers support a feature called  DLCI bundling to allow the splitting of one synchronous port with multiple DLCIs into a maximum of 32 Frame Relay subinterfaces. In this configuration, the DLCIs are divided between the subinterfaces through the use of the location table and the DLCI table. Each subinterface must have its own subnet or assigned network. The PortMaster has a limit of 512 total active interfaces, which can be further limited by available memory.

  The port you are configuring must be set for network hardwired use and Frame Relay, and must be in the same dial group as the location.

       Configuring Subinterfaces

  The following sections describe how to configure a Frame Relay subinterface.

 

Adding a Location

  To configure a Frame Relay subinterface, you add a location for each interface, configure it with the frame protocol, and associate it with a dial group. Then associate a synchronous port with the same dial group. For example, to create a location called sub1 , enter the following commands:

  Command> add location sub1
 Command> set location sub1 protocol frame
 Command> set location sub1 group 1
 Command> set W1 group 1 

  The rest of the location table entries are set as described in Chapter 7, "Configuring Dial-Out Connections," including setting an IP address, routing, and filtering for each interface.

 

Creating a DLCI Entry

  The next step in configuring the subinterfaces is to create an entry in the DLCI table. Entries can be followed with an optional IP address or hostname. The keyword ipxdlci  is available for IPX networks.

  Note ¯ The PortMaster 4 supports the IPX protocol if its running ComOS 4.1 or later. ComOS 4.0 does not support the IPX protocol.

  To create a DLCI table entry for the subinterface sub1 , enter the following commands:

  Command> add ipdlci sub1 16
 Command> add ipdlci sub1 19 192.168.2.19
 Command> add ipdlci sub1 20 192.168.2.20
 Command> add ipxdlci sub1 21 0e0a001e 

  To remove an entry, enter the following commands:

  Command> delete dlci sub1
 Command> delete ipxdlci sub1 21 

 

Displaying DLCI Entries

  DLCI entries that are added or deleted are linked to the location table. Use the show location  Locname command to display the DLCI entries.

       Troubleshooting Subinterfaces

  Packets received on a subinterface can be identified as belonging to that subinterface only if the DLCI is properly entered in the DLCI table for that location. If you are having problems, do the following:

•  Wait a few moments. Subinterfaces come up after the primary interface. This process can take a few moments.
•  Check the list of DLCIs tied to each location using the show location  Locname command.
•  Verify the DLCI list on a location using the show arp  Interface  command, replacing Interface with the name of the interface. A list of interfaces can be shown with the ifconfig  command.
•  Always reset the port after changing the DLCI list.
•  Verify that all DLCIs are accounted for by checking the DLCI list for your primary interface. If you enter the wrong DLCI for the subinterface, the DLCI for the subinterface is applied to the primary interface if LMI or Annex-D is in use.
•  Enter the following two commands to view the LMI or Annex-D keepalives:

    Command> set console  C0
  Command> set debug 0x51 

    After you verify that the proper keepalives are being received, enter the following commands to turn off the debug utility:

    Command> set debug off
   Command> reset console 

•  If you have a Cisco router on the other end of your connection, verify that it is set for encapsulation  frame-relay ietf  for the serial interface; otherwise, the Cisco frame-relay map  command for your DLCI must have the ietf  keyword appended.