TOKEN-RING STANDARD : ALGORITHMS AND SIMULATION

Abstract

Token- Ring is one of the popular LAN standards known in computer networks. It has been originally developed by IBM, and was later accepted as a standard by IEEE in 1985 (IEEE 802.5 standard).The main characteristics of a Token - Ring are its deterministic access control algorithm. and its ability to allow prioritized operation, and provision of frame acknowledgment. However, the implementation of this powerful access scheme requires additional management tasks such as token generation, duplicate address checking, orphan frames handling, ..., and others. These tasks are carried out using special frames called Medium Access Control (MAC) frames.

This thesis presents an integrated study of the Token- Ring LAN including its structure, topology and its various frame formats. It also describes the detailed operation of its medium access control algorithm. and the mechanism for prioritized operation. The three basic modes of operation involved in the Token- Ring algorithm are also described, namely: normal mode, standby mode, and active mode.

The control logic of the network card in a Token-Ring LAN has been analyzed in details with the help of the documents for IEEE 802.5 standard. Referring to various finites- state machine diagrams defining the operation of the Token- Ring in its various modes of operation, a separate comprehensive flow chart has been deduced for each operational mode. These can serve as the basics for simulating the operation ofthe Token-Ring LAN considering the exchange of management (control) frames, along with the usual data frames. As an example of building the simulation model, the Token- Ring LAN operation in the normal mode has been considered, and a computer code was written in C to simulate the various events considering prioritized operation. Considering such simulation, various network performance measures, including throughput and delay, have been determined and their effect is evaluated.

The advantages of simulation models stem from the fact that they can predict network performance when assumptions underlying the analytical results are not satisfied.