A Cellular Automaton Hardware Design For Generating Traffic Distributions For ATM Networks

Abstract: Broadband-Integrated Services Digital Network (B-ISDN) is capable of supporting the transmission of multimedia applications such as data, voice and images concurrently, because it uses Asynchronous Transfer Mode (ATM) as the basis and ATM is a high-bandwidth, low-delay, connection-oriented cell switching and multiplexing technology. Traffic of different applications has different traffic characteristics. To evaluate the performance of ATM networks, traffic source models with a desired bandwidth to meet the requirements of high speed ATM networks are required. These models should also be accurate to simulate traffic existed in real networks. ATM traffic modeling is a rapidly developing research area. From traditional Markov process models [1][2] to recent self-similar traffic modeling [3][4][5], researchers are trying every approach (theoretical and practical) to model actual network traffic. Some common approaches for modeling traffic arrivals are Poisson process, exponential distribution, Bernoulli distribution and Markov process. However, modeling stochastic processes with various distributions is required no matter which approach is chosen. The point is to make the model suitable for simulating realistic traffic. A Cellular Automaton (CA) consists of a regular uniform array of nearest-neighbor interconnection with combinational logic. CA can effectively generate random patterns which have good randomness characteristics. This paper makes the first step to use One-Dimensional Linear Hybrid CA (1-D LHCA) to generate random traffic distributions for ATM networks. The paper is organized into three sections: Section I is an introduction to CA and ATM traffic source models; Section II develops a CA model for a random sequence generator, presents the hardware implementation, simulation results and comparison between different schemes; and Section III is conclusions.