SuperWeb: towards a global Web-based parallel computing infrastructure

Abstract

The Internet, best known by most users as the World-WideWeb, continues to expand at an amazing pace. We propose a new infrastructure to harness the combined resources, such as CPU cycles or disk storage, and make them available to everyone interested. This infrastructure has the potential for solving parallel supercomputing applications involving thousands of cooperating components. Our approach is based on recent advances in Internet connectivity and the implementation of safe distributed computing embodied in languages such as Java.We developed a prototype of a global computing infrastructure, called SuperWeb, that consists of hosts, brokers and clients. Hosts register a fraction of their computing resources (CPU time, memory, bandwidth, disk space) with resource brokers. Client computations are then mapped by the broker onto the registered resources. We examine an economic model for trading computing resources, and discuss several technical challenges associated with such a global computing environment.

Browser-based Harnessing of Voluntary Computational Power

“…Javelin is a prototype implementation of the SuperWeb architecture [2]. The architecture consists from brokers, hosts and clients and, in general, does not depend upon Java/browser combination.…”

“…For example, they envisaged that when scientists would required more resources to carry the tests for their paper to be submitted to a conference with an approaching deadline, they would be willing to pay more for the resources [2].…”

“…To attract new users, a flash widget used by OGC [61] allowed to inspect the news fetched from Yahoo Top Stories feed, and in theory could be used to allow social interaction or playing games. Users could run parasitic games, paying with their CPU time by allowing some computation job running in background [2]. This concept was tested in [50] where users played simple snake game, while in the background workers were doing useful work.…”

“…In [2] it was suggested that program specific checksum could be computed by tasks to indicate they were actually run. To compute just and correct credit amount assigned to users after computation, small microbenchmarks might be first sent to a user to estimate his/hers machine computing capabilities [83].…”

“…The jobs are accepted only from trusted institutions, tested first on set of dedicated users, and finally uses pair of public/private keys is used to ensure code integrity [28]. Unfortunately, sometimes data in distributed computation could be sensitive, and would have to be encrypted while residing on hosts' side [2]. Similarly, application code could be obfuscated [68] [8].…”