Fig. 1. An example SIP call.

Fig. 2. Client-based failover.

Fig. 3. DNS-based failover.

Fig. 4. Failover based on database replication.

Fig. 5. When the primary server fails.

Fig. 6. When the master database fails.

Fig. 7. Co-located database and proxy.

Fig. 8. Reliable server pooling for SIP.

Fig. 9. Failover in CINEMA.

Fig. 10. DNS-based.

Fig. 11. Identifier-based load sharing.

Fig. 12. Two-stage reliable and scalable architecture.

Fig. 13. Example test setup for S3P3.

Fig. 14. Example message flow.

Fig. 15. Server throughput in S_n P_m configuration (n first stage and m second stage servers). The results show that the performance increases linearly with the number of servers, i.e., s2p2 is twice and s3p3 is thrice that of s1p1 and s0p1 performance.

Fig. 16. Theoretical and experimental capacity for configuration S_n P_m.

Fig. 17. Effect of user identifier distribution among second stage servers for S2P2. Uniform distribution gives the best performance, i.e., success rate is close to 100% until the peak performance (1800 CPS), whereas for non-uniform distribution the success rate reduces as soon as one of the server is overloaded (at 1500 CPS).

Fig. 18. Performance of S_n P_m with stateful proxy in second stage. The results show that the performance increases linearly with the number of servers, i.e., s2p2 is twice and s3p3 is thrice that of s1p1 and s0p1 performance.

Fig. 19. Stateful proxy message flow.

Fig. 20. Performance for S_n P_m with registration server in second stage. The results show that the performance increases linearly with the number of servers, i.e., s2p2 is twice and s3p3 is thrice that of s1p1 and s0p1 performance.

Fig. 21. REGISTER message flow.

Fig. 22. Processing steps in a SIP server.

Fig. 23. Performance of software architectures relative to event-based on different hardware for (a) stateless proxy, and (b) stateful proxy mode in SIP. For example, the performance of stateless proxy on 4 × P hardware in the thread pool architecture is approximately three times that in the event-based architecture on the same hardware.

Performance (CPS) of stateless proxy for Proxy 200 test

Performance (CPS) for stateful proxy for Proxy 200 test