GMaker: A video recommendation module for peer-assisted VoD - Springer

摘要

Peer-assisted Video-on-Demand (VoD) systems allow users to watch different parts of a video at the same time, which leads to an inefficient video content sharing among the users. Therefore, in P2P-VoD, it is difficult to alleviate the server load and meanwhile maintain a smooth playback. Nevertheless, if the population of a swarm is sufficiently large and under the server’s capacity, the efficiency can be improved to some extent. We, therefore, introduce the concept of Groupon into P2P-VoD system. Groupon is the most successful web site that uses the business model called deal-of-the-day that can attract a massive group of people and thus benefit both the customers and the retailers. Thus, a Groupon-like module named GMaker is proposed in P2P-VoD, which recommends the most suitable video to the user. In this case, a moderate number of audiences are gathered in one overlay. As a result, the upload capacity of the user is sufficiently used and the workload of media server can be alleviated. We support our design with extensive simulations to compare the performance of P2P-VoD applying GMaker and the one not applying GMaker in different scenarios. It shows that GMaker significantly enhances the performance of P2P-VoD resource utilization, quality of service of users and video support capacity.