Separating Performance Anomalies from Workload-Explained Failures in Streaming Servers
CCGRID '12 Proceedings of the 2012 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (ccgrid 2012)
| Hi-index | 0.00 |
With the advances of compression technology, storage devices and networks, Video on Demand (VoD) service is becoming popular. The system needs to provide continuous service and heterogeneous service levels for users. However, these requirements cannot be satisfied in conventional VoD system which is constructed on redundant content servers and centralized management. In this paper, autonomous VoD system is proposed to meet the requirements. The system is constructed on Faded Information Field architecture. Under the proposed architecture, autonomous fault detection and fault recovery technologies are proposed to achieve fault-tolerance for continuous service. The effectiveness of the proposed technologies are proved through simulation. The results show that an average of 30% improvement in recovery time and users' video service can be recovered without stopping compared with conventional VoD system.