Frame Relay is one of the most commonly used Layer 2 technologies.
Any WAN you ever work with is most likely going to have Frame Relay in there somewhere, and you’ve got to know how to work with it and how to troubleshoot it.
I advise you to practice Frame Relay in a lab environment, especially the Frame Relay debugs.
Chris Bryant CCIE #12933
Commands Used In This Chapter And Labs Include:
Debug frame lmi -- Used to detect LMI type a router is using.
Mismatched LMIs lead to Frame Relay failures.
Encapsulation frame relay – Enables Frame Relay encapsulation on an interface.
<no> frame-relay inverse-arp – Frame Relay ARP performs a dynamic DLCI – IP Address mapping function that may not always be desired. To prevent dynamic mapping, run the no frame-relay
inverse-arp command.
Frame-relay map – Used to configure manual DLCI – IP Address statements.
Frame-relay lmi-type – Used to manually configure an interface’s LMI type.
Show frame-relay lmi – Displays number of LMIs that have been received and sent, and status timeouts that may lead to a Frame Relay connectivity issue.
Show frame-relay map – Displays dynamic and static DLCI – IP address mappings.
Show frame-relay pvc – Displays locally configured DLCIs, how long they have been up, and the last time the PVC status changed.
Frame Relay is a Layer 2 connectivity method that delivers one major benefit over point-to-point links: cost.
Frame Relay uses virtual circuits, either permanent or on-demand, which is a logical path between two DTEs. There will not be a direct physical connection between the two DTEs. Instead, a logical
connection is formed through a path of DCEs. This path is referred to as a frame relay cloud.
Many users will share the same frame relay cloud. The frame relay service provider guarantees a certain amount of bandwidth will be available to a given user at any time. The more guaranteed bandwidth desired, the more it costs, but it’s still cheaper than a dedicated point-to-point link. This guaranteed bandwidth is referred to as the
committed information rate (CIR).
There are two types of virtual circuits. A permanent virtual circuit (PVC) is available at all times, where a switched virtual circuit (SVC) is up only when certain criteria are met. The configuration of a Frame Relay SVC is beyond the scope of the CCNA exam and will not be discussed here.
In The REAL World…
Frame Relay SVCs are used in production networks, but they’re rare.
Why? Because PVCs are so cheap. It’s important to know SVCs exist, though.
A PVC can be full-mesh or partial-mesh. A full mesh describes a topology where every router has a logical connection to every other router in the frame relay network.
More common is the partial-mesh configuration, where a single router (the hub) has a logical connection to every other router (the spokes), but the spokes do not have a logical connection to each other.
Communication between spokes will go through the hub.
The LMI: The Heartbeat Of Frame Relay
The Local Management Interface (LMI) messages are sent between the DCE, typically the service provider, and the DTE, the Cisco router. LMI Status messages serve as keepalives for the frame connection. If keepalives are not continually received by both the DCE and DTE, the frame connection will drop. The LMI also indicates the PVC status to the router, reflected as either active or inactive.
The LMI types must match on the DTE and DCE for the PVC to be established. There are three types of LMI: Cisco, ansi, and q933a.
Cisco routers feature LMI Autosense, where the router will send out an LMI Status message for all three LMI types. The router then waits for a response for one of those LMI types from the DCE, and then sends out its own LMI message to match that LMI type.
In The REAL World…
LMI Autosense sounds great, doesn’t it? Just plug and play. Except for one small detail: it doesn’t always work. Get in the habit now of hard-coding your LMI types. This skill will come in handy as you progress up the Cisco certification ladder, and on the job as well.
Encapsulation Types
Where LMI types must match between the DCE and DTE, the Frame encapsulation type must match only between the DTEs. The DCEs do not care about the encapsulation type. VCs on the same router can use different encapsulation types. The Cisco-proprietary encapsulation type is cisco, and the other choice is ietf. The default type is cisco, but this is not listed in IOS help:
R1#conf t
R1(config)#interface serial 1
R1(config-if)#encapsulation frame-relay ? ietf Use RFC1490/RFC2427 encapsulation
The default frame relay encapsulation type, cisco, is not listed in IOS Help.
Keeping track of which devices must agree on what can be a little confusing at first. This chart sums it up:
Topic Choices Who Must Agree
Encapsulation
Type CISCO, IETF Both Routers
LMI Type CISCO, ANSI,
Q933A Both Routers and their Neighboring Frame Relay Switch.
DLCI Addressing
Frame Relay VCs use Data-Link Connection Identifiers (DLCIs) as their addresses. Unlike other Cisco technologies, VCs have only a single DLCI in their header. They do not have a source and destination.
The reason is that DLCIs have local significance only. The same DLCI can be used on different access links in the same network, but the same DLCI cannot be used on multiple access links on the same router.
Local Significance vs. Global Addressing
Cisco uses the term global addressing to describe a technique by which a router in a Frame network is reached via the same DLCI number from each router in the network. For example, in a 250-router
network, the same DLCI number would be used to reach “Router A” by each router.
Global Addressing is an organizational tool that does not affect the fact that DLCIs have local significance only.
DLCI – to – IP Address Mapping
The locally significant DLCI must be mapped to the destination router’s IP address. There are two options for this, Inverse ARP and static mapping.
What is ARP? What is Inverse ARP?
Address Resolution Protocol (ARP) is used by a router when the router knows the Layer 3 address (the IP address), but not the Layer 2
address (the MAC address. ARP dynamically maps that Layer 3 address to the Layer 2 address.
Inverse ARP performs the opposite function. In this case, the Layer 2 DLCI address is known, but the Layer 3 IP address is unknown. The Layer 3 addresses are learned once the PVC is up. The DLCI-IP address mapping then takes place.
In both the following examples, the single physical Serial interface on Router 1 is configured with two logical connections through the frame relay cloud, one to Router 2 and one to Router 3.
Inverse ARP runs by default once Frame Relay is enabled. By running show frame-relay map after enabling Frame Relay, two dynamic mappings are shown on this router. If a dynamic mapping is shown, Inverse ARP performed it.
Dynamic frame mappings are performed by Inverse ARP, enabled by default on a frame relay interface.
R1#show frame map
Serial0 (up): ip 200.1.1.2 dlci 122(0x7A,0x1CA0), dynamic, broadcast,, status defined, active
Serial0 (up): ip 200.1.1.3 dlci 123(0x7B,0x1CB0), dynamic, broadcast,, status defined, active
Static mappings require the use of a frame map statement. To use static mappings, turn Inverse ARP off with the no frame-relay inverse-arp statement, and configure a frame map statement for each remote destination that maps the local DLCI to the remote IP address. Frame Relay requires the broadcast keyword to send broadcasts to the remote device.
R1#conf t
R1(config)#interface serial0
R1(config-if)#no frame-relay inverse-arp
R1(config-if)#frame map ip 200.1.1.2 122 broadcast R1(config-if)#frame map ip 200.1.1.3 123 broadcast
The syntax of the frame map statement maps the remote IP address to the local DLCI.
Broadcasts will not be transmitted by default; the broadcast option must be configured.
R1#show frame map
Serial0 (up): ip 200.1.1.2 dlci 122(0x7A,0x1CA0), static, broadcast,
CISCO, status defined, active
Serial0 (up): ip 200.1.1.3 dlci 123(0x7B,0x1CB0), static, broadcast,
CISCO, status defined, active
The “static” status of the mapping indicates it was configured manually.
In The REAL World…
Like LMI Autosense, Inverse ARP sounds like a great idea, and doesn’t always work well. You’re much better off disabling Inverse ARP
whenever possible and configuring static mappings. For exam purposes, know how to do both. As you progress up the Cisco
certification ladder toward the CCIE, you’ll find that disabling Inverse ARP becomes a habit.
Notes