By the rapid increase of the content-oriented applications, storing and distributing contents become more valuable. As a future of the Internet architecture, Information-Centric Networking (ICN) has a novel feature called in-network caching, to cache the content in ICN routers. It avoids delivery of the same content many times in the same path and reduces user-perceived delay, server load, waste of bandwidth, etc. Recently many algorithms were proposed for the cache placement strategy. One of the traditional cache placement policy is Leave Copy Everywhere (LCE). LCE caches a copy of data passing through an ICN router. LCE is not efficient because of cache redundancy. In this work, we propose a new algorithm called LCD-Based Probabilistic Caching (LBPC). When a data packet is delivered back to the requester, the first ICN router located closer to the server caches the content with the highest probability, and the cache probability at each router decreases along the path. When multiple requests are sent for the same content, the copy of the content is pushed to the downstream routers hop by hop. Our simulation work shows LBPC has better performance than In-Network Probabilistic Caching (ProbCache) LCE, and Leave Copy Down (LCD).
展开▼
