Previous | Table of Contents | Next |
Frame relays packet-switched technology is based on its older relative, X.25. In a packet-switched network, data is subdivided into individual packets, each with a unique identification and destination address. These packets, or frames, as they are called in frame relay, are variable in length. Each frame contains a header, information field, frame check sequence (FCS), and two flags. (See Exhibit 1.)
The header of a frame contains information on connection identification. Connection identification is managed by the data link connection identifier (DLCI). The DLCI is an identifier (i.e., address) associated with each permanent virtual circuit (PVC). A PVC is the path that is set up by the service provider routing data from point A to point B. Once a PVC is defined, it requires no setup operation before data is sent and no disconnect operation after data is sent.
The DLCI is usually 10 bits long, but it can be extended. Each DLCI indicates a different PVC or end point for data. In other words, the DLCI must be unique for specific destinations. For example, for San Francisco, the DLCI is 45; for New York, it is 49. These values ensure the proper routing of information within a network (see Exhibit 2). The DLCI can be extended to increase addressing capabilities. There are two, one-bit Extended Address fields in the header that indicate if another octet has been added for extended DLCI purposes.
Exhibit 2. Sample DLCI Values.
Also within the header are congestion control bits, which identify whether congestion is present. These congestion bits are important because they inform the user of potential errors in the network. If congestion overloads a particular switch, frames will be discarded by the network. There are three main types of congestion control bits: forward explicit congestion notification (FECN), backward explicit congestion notification (BECN), and discard eligibility (DE). (See Exhibit 3.)
The header also includes the Command/Response field. This one-bit field is used by many HDLC-based protocols to indicate whether a frame carries a command or a response. It is passed transparently through the Frame relay network.
The information field contains variable numbers of octets (up to 4,096 octets for some implementations) and encapsulates many protocols, including TCP/IP, internetwork packet exchange (IPX), sequenced packet exchange (SPX), SNA, and X.25. The information field is the largest part of the frame and contains the encapsulated protocol header as well as user data.
Exhibit 3. Congestion Control.
The FCS field is used for error detection. The transmitting end devices (i.e., routers) apply a complex algorithm to the data in the frame and then place the result in this field as the frame is sent to the receiving equipment (i.e., the router at the far end). Each node along the way recalculates the FCS based on the received data and compares the result to the FCS within the frame. If the results are not identical, then the frame is discarded by the switching node or the end device.
Flags serve as eight-bit idle codes that delineate frames on the circuit.
The local management interface (LMI), a signaling protocol, was instituted by Cisco Systems, Inc. (Menlo Park, CA), Digital Equipment Corp. (Maynard, MA), Northern Telecom, Inc. (McLean, VA) and StrataCom, Inc. (San Jose, CA). It allows the subscriber and the network devices to share information on the status of DLCI on the link.
In addition, the LMI provides a mechanism for the network to recognize that end devices are operating correctly. The LMI proposal was accepted with minor additions by the ANSI and the ITU. These standards organizations slightly modified and standardized the LMI, which is now represented by ANSI T1.617, Annex D, and ITU Q.933, Annex A.
Frame relay networks sacrifice function for speed by relying on intelligent devices and an efficient network to prevent errors and retransmissions. With such a reliance on the network, it is imperative that users test the physical medium, the PVC, and the internetwork, both out of service and in service, when implementing frame relay. Specifically:
The testing of the transmission facilities, as well as testing the internetwork service, ensures that the service itself, the internetwork devices, and the physical media provide seamless network performance for customers.
Users need to carefully test the transmission facilities before implementing a Frame relay network. There are three parts to a comprehensive test strategy: testing the transmission facilities between the subscriber and the POP switch, testing the PVC through the network, and conducting a lost frames analysis.
Bit Error Rate Testing. To test the transmission facilities between the subscriber and the POP switch, an out-of-service BER test should be performed. This test examines the transmission facilities by evaluating the percentage of bits received in error compared to the total number of bits received.
Typically, the BER test set generates one or more complex sets of pseudorandom bit patterns and then transmits these patterns along network segments at rates dependent on the facilities. The BER test set provides the stimulus as well as the receiver to monitor the response of the transmission segment being tested. This test can pick up bit errors, bipolar violations (BPV), cycle redundancy check (CRC) errors, and other errors that may cause the network to drop frames. When these errors occur, an end device requests retransmission, which costs money and consumes valuable bandwidth.
Because the pseudorandom patterns used in BER test do not pass through the switches, this test is useful only between the customer and the POP switch. However, because of frame relays lack of error correction utilities, BER testing is vital. A clean transmission line must be ensured before any additional investigation can be done.
PVC Testing. Once the transmission facilities are verified, the PVC should be tested. This test verifies end-to-end connectivity between two or more subscribers on a specific PVC,
This test is performed by generating and receiving a variety of Frame relay traffic loads with the proper DLCI and LMI support over each valid PVC. The test should be able to transmit frames with control of various header bits, like DLCI, FECN, BECN, and DES test results indicate that the transmission facilities are operating with minimal errors and the correct link management parameters and that the network is properly routing frames across the PVC being tested.
When testing transmission facilities, PVC, and LMI, monitoring of layer-two and layer-one (i.e., transmission facility and PVC phenomena is also important. For example, users can determine that a bipolar violation (BPV) at the physical layer (layer one) coincides with a FCS error on a layer-two analysis, which indicates a transmission facilities problem and not a Frame relay problem.
A Lost Frames Analysis. Another important feature of Frame relay testing is determining if any transmitted frames are lost. This test is performed by placing a sequential counter in the data field of the test frames. The test set receiver tracks gaps in the test frames sequence count. There are usually several reasons the network may be dropping frames: FCS errors are occurring, the attempted throughput exceeds the CIR, too many frames per second are being transmitted, or the network may be experiencing congestion.
This test also discovers dribbling errors that cause network bottlenecks and the determination of critical thresholds of Frame relay devices. The lost frame analysis gives customers a final assurance when verifying network performance because it can discover errors that previous tests did not catch.
Previous | Table of Contents | Next |