This Topic is devoted to High-Performance Computing on widely distributed systems utilizing Peer-to-Peer technologies. In contrast with the usual Client-Server organization scheme, these systems are (or attempt at being) highly decentralized, self-organizing with respect to the external context, and with a notion of balanced resource trading. Hence, each node within such a system can provide both Client and Server capability. In the long term, participating nodes may expect to be granted as much resources as they provide.
Because of size, autonomy and high volatility of their resources, P2P platforms provide the opportunity for researchers to re-evaluate many fields of Distributed Computing, such as protocols, infrastructures, security, certification, fault tolerance, scheduling, performance, etc.
This Topic focuses on High-Performance Computing in the broadest sense: putting together CPU cycles, acquiring input data, managing temporary and persistent data, interacting with the user, storing result files, etc. Work addressing new issues raised by the current trend for the convergence of Grid and P2P systems will be of particular interest: concerning installation, utilization, flexibility, scalability, etc.
Authors wishing to demonstrate running systems and applications in addition to their research paper are invited to refer to the specific Demonstration Topic 19.
Prof. Luc Bougé
IRISA, ENS Cachan, Brittany extension
Dr. Franck Cappello,
LRI-Université Paris-Sud, France