Harmonic multicast-based proxy prefix cache allocation scheme for video-on-demand services

Abstract

It is essential to reduce the server load for video-on-demand service, because many service requests and delivering large size of contents causes a big server load problem. The server-proxy model is e??ective to resolve the server load problem of Video-on-Demand service. In the server-proxy model, the limited storage of proxy server is an issue to reduce the server cost. To solve the issue, this thesis proposes the prefix cache allocation with multicast to minimize the bandwidth and storage costs which is determined by the video request rate. To optimize cache size, the harmonic multicasting method is applied. The proposed scheme determines the size of prefix cache as the length of a segment of harmonic multicast method. The number of segments is calculated that minimize the bandwidth and storage costs.