Autonomous fault recovery technology for continuous service in Distributed VoD system

Abstract

In the video on demand (VoD) service, users request heterogeneous video quality based on their preference and usage environment which are dynamically changing. On the other hand, service providers request to provide variety service with minimum total storage volume in the system because the volume of the video data is too huge. In addition, considering the characteristics of the application, services must be distributed without stopping playback or deteriorating video quality. However, conventional VoD system constructed on redundant content servers and centralized management cannot satisfy these requirements. Autonomous VoD system (AVoDS) is proposed to meet these requirements. This system is based on the faded information field (FIF) architecture in which each node collaborate with the other nodes for service provision and utilization supported by the mobile agent. In this system, the layered streaming video data are employed and each data are distributed on different service layers. Therefore the system can provide service with adaptability and minimum total storage volume. In this paper, implementation of the prototype of AVoDS is introduced and autonomous fault recovery technology is proposed for the continuous service provision. The fault of the node is autonomously detected by the connected nodes and the fault recovery processes for each user are widely distributed to other nodes in the system. The effectiveness of the proposed technology is proved through simulation