作者: Ying Cai , Zhan Chen , Wallapak Tavanapong
DOI: 10.1007/S11042-007-0136-5
关键词: Peer-to-peer 、 Smart Cache 、 Distributed computing 、 Cache algorithms 、 Host (network) 、 Cache invalidation 、 Computer science 、 Cache 、 Scalability
摘要: Providing scalable video services in a peer-to-peer (P2P) environment is challenging. Since videos are typically large and require high communication bandwidth for delivery, many peers may be unwilling to cache them whole serve others. In this paper, we address two fundamental research problems providing P2P services: (1) how host can find enough pieces, which scatter among the system, assemble complete video; (2) given limited buffer size, what part of should existing data expunged make necessary space. We these with new ideas: Cell caching collaboration Controlled Inverse Proportional (CIP) allocation. The concept allows cost-effective fully distributed dramatically reduce lookup cost. On other hand, CIP allocation challenges conventional wisdom by unpopular higher priority. Our approach system retain copies popular avoid creating hot spots at same time, prevent from being quickly evicted system. have implemented Gnutella-like simulation network use it as testbed evaluate proposed technique. extensive study shows convincingly performance advantage scheme.