| Summary: | 碩士 === 國立臺北科技大學 === 資訊工程系研究所 === 101 === Peer to Peer (P2P) technique over internet had been developed for years. Currently, watching TV shows, movies, or even the live video through P2P-based applications are very common things. P2P applications can be divided to four types: file sharing systems (e.g. Bit Torrent), instant messaging systems (e.g. GTalk), media streaming systems (e.g. PPLive), and storage sharing system (e.g. Grannary). One of the most important characteristics in P2P technique is sharing data efficiently, every client can share data to each other instead of getting data from the same source, and this technique makes the resource get higher usage.
This thesis realized two ALM algorithms to deal with issues of recovery time and multisource. The Distributed Multisource Forwarding Trees (DMFT) and Cluster-based Distributed Multisource Forwarding Trees (CDMFT) were implemented and tested. The key performance measured is the recovery time for the delivery of multicast media data when a group member in a multicast group either joins or leaves the group.
|
|---|
