P2P-based Job Assignment for Public Resource Computing

Complex applications often require the execution of a large number of jobs in a distributed environment. One highly successful and low cost mechanism for acquiring the necessary compute power is the "public resource computing" paradigm, which exploits the computational power of private computers. However, applications that are based on this paradigm currently rely upon centralized job assignment mechanisms that can hinder the achievement of performance requirements in terms of overall execution time, load balancing, fault-tolerance, reliability of execution results, scalability and so on. This paper extends a superpeer protocol, proposed earlier by this group, for the execution of jobs based upon the volunteer requests of workers. The paper introduces a distributed algorithm that aims to achieve a more efficient and fair distribution of jobs to workers. This is obtained by the definition of different roles that can be assumed by super-peers and ordinary nodes on the basis of their characteristics. A simulation study is carried out to analyze the performance of the super-peer protocol and demonstrate the advantage of distributing the job assignment process.