Simulation methods for analysis of traffic processes in ATM networks

Abstract

This paper presents efficient simulation methods for analyzing modern, large-scale networks and evaluating their performance attributes. Characterizing traffic flows from multiple sources and applications is key in assessing overall network performance measures. It is essential to have quantitative network cost and performance measures in order to plan, design, and implement modern, large-scale networks such as the Advanced Distributed Learning Network (ADLN). ADLN requires integrated, multimedia network services for distributed, collaborative processing among globally dispersed users. ADLN will interconnect multiple categories of users and provide integrated voice, video, and data services, which can be enabled through the use of Asynchronous Transfer Mode (ATM) technology (ATM Forum 1992). Through efficient multi-plexing and networking, ATM can interconnect multiple classes of users and transport ADLN applications cost-effectively with guaranteed per-formance. A combined approach of simulation and analysis is used to assess the performance of the large-scale, distributed network such as ADLN. Topics include the multi-source traffic characterization, performance analysis of ATM networks, capacity sizing, and optimal allocation of network resources.