A P2P Music-on-Demand Protocol and System Using Scalable Audio Codec

• Main Authors: Wei-Chung Chou, 周唯中
• Other Authors: Alvin Wen-Yu Su
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/62340433617704027839

碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 96 === SNP （SCREAM Network Platform） is an audio-on-demand platform developed by SCREAM Lab. The main purpose of SNP is to provide fault-tolerant, scalable, load balancing and QoS streaming service. SNP adopts the ideas of client-server and p2p network to build its architecture. The internal mechanism can divide to local and global by locality. The need of computation and bandwidth is ordinarily distributed amount network and these mechanisms are completed by peer nodes. During turbulent period (e.g. the number of available peers is little), these mechanisms will takeover by server nodes. In this cooperative model, SNP can keep the advantage of distribution but also reliability. To store the audio files, SNP views peer nodes as its repository. The audio files are encoded by scalable multi-layered codec. When deployment, SNP slice single file into base layer, enhancement layer 1, enhancement layer2 … The most important layer will be stored in most powerful peer node and so on.