A Markov Model for Preference List Sizes in Combined Input Output Queued Switches with Stable Matching Algorithms

Abstract

Combined Input Output Queued switches have emerged as one solution to overcome the speedup problem in pure output queued switches. By using a stable matching algorithm, a speedup of two is sufficient to exactly mimic the behavior of an output queued switch. The size of preference lists is a key parameter in the stable matching algorithm that ultimately affects the algorithm runtime and complexity. A model using Markov chains is presented to predict the preference list sizes for uniform traffic. The model achieves high accuracy across a wide range of switch sizes and traffic loads.