A Queueing Theoretic Model for Service Discovery in Ad-hoc Networks

Abstract

This paper describes an analytical model for service discovery protocols in ad-hoc networks. We propose a bottom-up approach towards modeling various essential features of distributed service discovery protocols. More specifically this paper proposes a method to analytically model the service cache (SC) on an ad-hoc node. %A Service Cache stores service advertisements. We employ the use of stochastic process models and fundamental queuing theory to model the data present in the cache. We show using simulations how our model can be used to predict the service cache usage at the node level. Our model is based on the assumption of non-preemption of service descriptions from the service cache and that the system is at steady state. We assume that there are infinite number of services in the system. We also experimentally verify that our model predicts the service system behavior reasonably well even when the assumption of infinite services does not hold.