An Algorithmic Framework for Discrete-Time Flow-Level Simulation of Data Networks

摘要

In this paper, we present a comprehensive algorithmic framework for discrete-time flow-level simulation of data networks. We first provide a simple algorithm based upon iterative equations useful for the simulation of networks with static traffic demands, and we show how to determine packet loss and throughput rates using a simple example network. We then extend these basic equations to a simulation method capable of handling queue and link delays in dynamic traffic scenarios and compare results from flow-level simulation to those obtained by packet-level simulation. Finally, we illustrate the tradeoff between computational complexity and simulation accuracy which is controlled by the duration of a single iteration interval Δ.