Performance Analysis of FDDI

The performance of an FDDI LAN depends upon configuration and workload parameters such as the extent of the ring, the number of stations on the ring, the number of stations that are waiting to transmit, and the frame size. In addition, one key parameter that network managers can control to improve performance is the target token rotation time (TTRT). Analytical modeling and simulation methods were used to investigate the effect of the TTRT on various performance metrics for different ring configurations. This analysis demonstrated that setting the TTRT at 8 milliseconds provides good performance over a wide range of configurations and workloads. Fiber distributed data interface (FDDI) is a 100megabit-per-second (Mb/s) local area network (LAN) defined by the American National Standards Institute (ANSI).[1,2] This standard allows as many as 500 stations to communicate by means of fiber-optic cables using a timed-token access protocol. Normal data traffic and time-constrained traffic, such as voice, video, and real-time applications, are supported. All major computer and communications vendors and integrated circuit manufacturers offer products that comply with this standard. Unlike the token access protocol defined by the IEEE 802.5 standard, the timed-token access protocol used by FDDI allows synchronous and asynchronous traffic simultaneously.[3] The maximum access delay, i.e., the time between successive transmission opportunities for a station, is bounded for both types of traffic. Although this delay is short for synchronous traffic, that for asynchronous traffic varies with the network configuration and load and can be as long as 165 seconds. Long maximum access delays are undesirable and can be avoided by properly setting the network parameters and configurations. This paper begins with a description of the timedtoken access method used by FDDI stations and then proceeds to discuss how various parameters affect the performance of these systems. The target token rotation time (TTRT) is one of the key parameters. We investigated the effect of the TTRT on FDDI LAN performance and developed guidelines for setting the value of this parameter. The results of our investigation constitute a significant portion of this paper. Timed-Token Access Method The token access method for network communication, as defined by the IEEE 802.5 standard, operates in the following manner. A token circulates around the ring network. A station that wants to transmit information waits for the arrival of the token. Upon receiving the token, the station can transmit for a fixed time interval called the token holding time (THT). The station releases the token either immediately after completing transmission or after the arrival of all the transmitted frames. The time interval between two successive receptions of the token by a station is called the token rotation time (TRT). Using this scheme, a station on an nstation ring may have to wait as long as n times the THT interval to receive a token. This maximum access delay may be unacceptable for some applications if the value of either n or THT is large. For example, voice traffic and real-time applications may require that this delay be limited to 10 to 20 milliseconds (ms). Consequently, using the token access method severely restricts the number of stations on a ring. The timed-token access method, invented by Grow, solves this problem by ensuring that all stations on a ring agree to a target token rotation time (TTRT) and limit their transmissions to meet this target.[4] There are two modes of transmission: synchronous Digital Technical Journal Vol. 3 No. 3 Summer 1991 1 Performance Analysis of FDDI and asynchronous. Time-constrained applications such as voice and real-time traffic use the synchronous mode. Traffic that does not have time constraints uses the asynchronous mode. A station can transmit synchronous traffic whenever it receives a token; however, the total transmission time for each opportunity is short. This time is allocated at the ring initialization. A station can transmit asynchronous traffic only if the TRT is less than the TTRT. The basic algorithm for asynchronous traffic is as follows. All stations on a ring agree on a target token rotation time. Each station measures the time elapsed since last receiving the token, i.e., the TRT. On token arrival, a station that wants to transmit computes a token holding time using the following formula: THT = TTRT TRT If the value of THT is positive, the station can transmit for this time interval. After completing transmission, the station releases the token. If a station does not use its entire THT, other stations on the ring can use the remaining time through successive applications of the algorithm. Note that even though the stations attempt to keep the TRT below the target, they do not always achieve this goal. The TRT can exceed the target by as much as the sum of all synchronous-transmission time allocations; however, these allocations are limited so that their sum is less than the TTRT. As a result, the TRT is always less than twice the TTRT. Performance Parameters The performance of any system depends upon both system parameters and workload parameters as shown in Figure 1. There are two kinds of system parameters: fixed and user-settable. Fixed parameters cannot be controlled by the network manager and vary from one ring to another. Cable length and the number of stations on a ring are examples of fixed parameters. It is important to study network performance with respect to these parameters; if performance is sensitive to them, each set of fixed parameters may require a different guideline. System parameters that can be set by the network manager or the individual station manager include various timer values. Most of these timers influence the reliability of the ring and the time it takes to detect a malfunction. The key settable parameters that affect system performance are the TTRT and the synchronous time allocations. PERFORMANCE PARAMETERS