A novel PQPR algorithm for SIP signaling race detection

Session Initiation Protocol (SIP) is a significant protocol in Voice over Internet Protocol (VoIP) network to create, manage, control and tear down the sessions. During multimedia conversation, overlapping of signaling message occurs that is referred as SIP race condition which directly affects voice call quality. The existing solution is adjusting session timer would minimize race conditions, but media conversation phase is not constant over the time period, thereby not being a feasible solution for such conditions. In this paper, we propose a novel Priority Queue based Packet Rotating (PQPR) algorithm and design a handler to detect and minimize the signaling race packet before serving to the server thereby improving the voice call quality. This paper presents SIP signaling race condition and analyzed using an experimental test bed. The results of the experiment show that PQPR algorithm achieves better throughput with less than 3 % of false positive with minimal packet dropping.