Parallel processing over a peer-to-peer network constructing the poor man's supercomputer.

Abstract

An enticing eventuality of any promising novelty is its evolution from an esoteric research area to an unbounded opportunity for the masses. Quintessential technological phenomena include the Internet, the Web, and peer-to-peer file-sharing, among others. However, the global community appears to be still hesitant at tapping into the well of unharnessed potential offered by exploiting distributed computing. This is despite the fact that the aggregation of typical home computers through a peer-to-peer framework over the Internet would yield a virtual supercomputer of unmatched processing power, 95% of which is presently being left unutilized. Reasons for such hesitation include the lack of personal incentive for participants, and the high degree of expertise required from application developers. In this project, we strive to tackle the aforementioned obstacles by building a peer-to-peer system capable of deploying user-defined tasks onto the network for distributed execution. Users would only be expected to write standard concurrent code, and may rely on the system to transparently provide for optimal task distribution and fault tolerance guarantees, including pre-processing the code to achieve strong mobility during process migration. The dynamic behaviour of our system builds on an interplay between the structural architecture of the Peer Controller and the virtualization platform given by the Leverage library {our API}. The functionality of the latter is enabled through the invocation of the former; however, it is exposed to the user at a substantial level of abstraction, since several mechanisms - such as task distribution, migration, message delivery, global state, fault tolerance, and recovery - would be provided by the underlying system.