Overview of Frame Relay

Frame Relay is a WAN service optimized for modern data applications. It has higher throughput than X.25, and it makes more efficient use of network resources than individual leased lines running basic HDLC.

Frame Relay more efficiently utilizes bandwidth of a physical link by allowing connectivity to multiple sites through a single physical connection. LAN traffic is bursty by nature, and Frame Relay takes advantage of that burstiness by layering multiple logical connections on a single physical circuit. The individual connections within the physical connection are known as virtual circuits (see Figure 16-1).

Figure 16-1. Typical Frame Relay Network

Frame Relay performance management consists of monitoring the utilization of a single physical link and the throughput and latency of the individual virtual circuits. Frame Relay fault management consists of monitoring for dropped or errored frames across the Frame Relay cloud.

Network managers have discovered that there are many Frame Relay vendors, and each has its own implementations and naming conventions. The terms discussed are described as they should work, but your actual performance may vary according to the service provider selected. However, the generic description of the following terms should make some sense out of what can be chaos:

• PVC (Permanent Virtual Circuit) is a Frame Relay logical connection. The PVC is defined by its end points and class of service. It is identified at each end point by a Data Link Connection Identifier.

• DLCI (Data Link Connection Identifier) is a unique number assigned to a PVC at one of its end points. It is a Layer 2 network address that is analogous to a MAC address on a LAN. As seen in Figure 16-1, DLCIs are locally significant, meaning that they are unique only on a particular physical link.

• CIR (Committed Information Rate) is the rate in bits per second a service provider agrees to transfer across a given PVC. This rate is averaged over the Committed Rate Measurement interval (T_c). B_c (Committed Burst) is the maximum number of bits that the switch has been programmed to transfer during any Committed Rate Measurement Interval (T_c) without any policing. B_e (Excess Burst) is the maximum number of uncommitted bits that the switch will attempt to transfer beyond the CIR.

• DE (Discard Eligible) is a bit in the Frame Relay header. When set, it indicates that the frame is eligible for discard in the event of network congestion. The bit is set when a frame entering the switch is determined to be in excess of the CIR.

• FECN (Forward Explicit Congestion Notification) is a bit in the Frame Relay header. It is set by the network Frame Relay switch to notify the end station receiving the frame that the frame has been delivered through a congested path of the network.

• BECN (Backward Explicit Congestion Notification) is a bit in the Frame Relay header. It is set by the network Frame Relay switch to notify the source (sending) station that congestion exists in the path it is transmitting into.

To put some of these terms into perspective, consider the following. When data is transmitted across the physical link into the Frame Relay switch, it is transferred at the link access rate. The switch counts the incoming bits on a per-VC basis as B_c bits within time interval T_c. Any bits arriving in excess of the B_c limit are counted as B_e bits, and the frame containing these bits will have the DE bit set. Frames with the DE bit set are forwarded if there is no congestion detected in the network. After you go beyond the B_e limit, the switch discards new incoming frames.