Using simulation to explore distributed key-value stores for extreme-scale system services
SC '13 Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis
| Hi-index | 0.00 |
This paper focuses on fault tolerance of super-nodes in P2P-SIP systems. The large-scale environments such as P2P-SIP networks are characterized by high volatility (i.e. a high frequency of failures of super-nodes). Most fault-tolerant proposed solutions are only for physical defects. They do not take into account the timing faults that are very important for multimedia applications such as telephony. We propose HP2P-SIP which is a timing and physical fault tolerant approach based on a hierarchical approach for P2P-SIP systems. Using the Oversim simulator, we demonstrate the feasibility and the efficiency of HP2P-SIP. The obtained results show that our proposition reduces significantly the localization time of nodes, and increases the probability to find the called nodes. This optimization allows to improve the efficiency of applications that have a strong time constraints such as VoIP systems in dynamic P2P networks.