Peer-to-Peer Neighbor Selection Using Single and Multi-objective Population-Based Meta-heuristics

Abstract

Summary. Peer-to-peer (P2P) topology has significant influence on the performance, search efficiency and functionality, and scalability of the application. In this Chapter, we introduce the problem of neighbor selection in peer-to-peer networks using two population based meta-heuristics: Particle Swarm Optimization (PSO) algorithms and Genetic Algorithms (GAs). Both a single objective and a multi-objective problem are formulated, and then the P2P neighbor selection problem is defined. We present the neighbor selection strategy based on PSO and GA algorithm. Each particle encodes the upper half of the peer-connection matrix through the undirected graph, which reduces the search space dimension. We also discuss the characteristics of ergodicity during particle swarm searching process. We also illustrate the algorithm performance and trace its feasibility and effectiveness with the help of some examples.

Data Scheduling in Data Grids and Data Centers: A Short Taxonomy of Problems and Intelligent Resolution Techniques

“…The idea is that we could defined a virtual overlay on top of the Grid system by defining neighboring relations among computing sites and data hosts if computing sites contain replicas of data fragments for execution of a task assigned to the computing site. Then, we can formulate an optimization problem consisting in finding a subset of peer neighbors of the computing site from where to download/receive the data fragments [43]. The problem can be formally defined as follows.…”

“…The near-optimal resolution of this problem [43] can be used at the scheduling phase of selecting data hosts from where to get the data need for completion of the tasks in the batch.…”