ERPC: An Edge-Resources Based Framework to Reduce Bandwidth Cost in the Personal Cloud

摘要

Personal Cloud storage and file synchronization services, such as Dropbox, Google Drive, and Baidu Cloud, are increasingly prevalent within the Internet community. It is estimated that subscriptions of personal cloud storage are projected to hit 1.3 billion in 2017. In order to provide high rates of data retrieving, cloud providers require huge amounts of bandwidth. As an attempt to reduce their bandwidth cost and, at the same time, guarantee the quality of service, we propose a novel cloud framework based on distributed edge resources (i.e., voluntary peers in P2P Networks and edge servers in Content Delivery Networks). P2P technique is well considered as an efficient way to distribute contents, but not all of contents are applicable for it. Thus we first present an approach to select the contents which are more profitable to be placed in P2P networks. Considering the unreliability of P2P networks, we utilize CDN servers to dynamically cache the contents which can not be well served by P2P peers. The proposed caching algorithm takes the state of P2P peers into account and considers replacement and allocation policies simultaneously. According to trace-driven simulations, the ERPC achieves an impressive performance of saving bandwidth for cloud system and guaranteeing download rates for users.