Rate Allocation over Incentive based Mobile Peer-to-Peer System

• Main Authors: Chang-Yu Sui, 許仲佑
• Other Authors: Tein-Yaw Chung
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/92602850872545088981

碩士 === 元智大學 === 資訊工程學系 === 97 === Mobile Peer-to-Peer networks (MP2P) are expected to be widely studied in the near future, In MP2P, it is a great challenge to encourage resource sharing and prevent free-riding problem between mobile devices with limited resource. The problem becomes even complex when mobile nodes are multimode devices and use TCP for data exchange. This Thesis makes use of parallel TCP connection to explore full utility of bandwidth. For MP2P data exchange, we model the problem to be Incentive based Rate Allocation and Routing (IRAR) problem which is NP-hard. First, we exploit a joint combination of Lagrangian with α-rate allocation [31][32][33] called LG_α_RA to solve IRAR problem. Second, we propose the light-load and heuristic mechanism called Multiple Path Connection Prediction (MPCP) to reduce the complexity by α-rate allocation. From simulation result, on one hand, the performance of IRAR can be enhanced and achieve fairness and incentive by LG_α_RA. On the other hand, the result of rate allocation by LG_MPCP is much close to LG_α_RA.