Grid computing has recently emerged as a response to the growing demand for resources (processing power, storage, etc.) exhibited by scientific applications. However, as grid sizes increase, the need for self-organization and dynamic reconfigurations is becoming more and more important. Since such properties are exhibited by P2P systems, the convergence of grid computing and P2P computing seems natural. However, using P2P systems (usually running on the Internet) on a grid infrastructure (generally available as a federation of SAN-based clusters interconnected by high-bandwidth WANs) may raise the issue of the adequacy of the P2P communication mechanisms. Among the interesting properties of P2P systems is the volatility of peers which causes the need for integration of a service fault tolerance. And service Load balancing, As a solution, we proposed a mechanism of fault tolerance and model of Load balancing adapted to a grid P2P model, named SGRTE (Monitoring and Resource Management, Fault Tolerances and Load Balancing).DOI: 10.18495/comengapp.12.071084
展开▼
