Evaluating SIP proxy server performance

Abstract

SIP is a protocol of growing importance, with uses for VoIP, instant messaging, presence, and more. However, its performance is not well-studied or understood. In this paper we experimentally evaluate SIP proxy server performance using micro-benchmarks meant to capture common SIP proxy server scenarios. We use standard open-source SIP software such as OpenSER and SIPp, running on an IBM BladeCenter with Red Hat Enterprise Linux and Gigabit Ethernet connectivity. We show performance varies greatly depending on how the protocol is used. Depending on the configuration, throughput can vary from hundreds to thousands of operations per second. For example, we observe that the choice of stateless vs. stateful proxying, using TCP rather than UDP, or including MD5-based authentication can each affect performance by a factor of 2–4.