Implementing state distribution model in asterisk server

Abstract

Requests in SIP (Session Initiation Protocol) usually pass through several servers. Since resources are considerably consumed to keep the state, state distribution problem is raised among several servers where the goal is to increase throughput, accessibility of servers and reduce overload. In this paper, the optimisation problem was formulated and implemented as a distributed algorithm on Asterisk proxy server which is a commercial proxy server and evaluated with Spirent traffic production device. It results in a more scalable SIP server which dynamically determines the number of SIP requests for which server is stateful and assigns state management to the server which is inferior to it for the remaining demands. The current SIP servers are statically configured such that they are stateless or stateful and don't give optimal efficiency.