| Summary: | 碩士 === 東海大學 === 資訊工程學系 === 100 === Build and maintain a distributed systems to process the huge computing needs is a high complexity and high cost of the work. Volunteer computing technology to use a lot of volunteer computer parallel computing power to efficiently complete the calculation ,and this model have been many scientific projects proved to have very good results. Volunteer computing system performance depends primarily on the volunteer computer availability and reliability, this paper examines the impact of the differences of individual computers, the availability of the overall system performance. We propose that the clustering of volunteer computer management system, the use of spot in-checking technology to view the volunteer computer availability and clustering. We have developed a new clustering scheduling method, a group has an associated workunit are allocated to the same group of volunteer computers, in order to reduce the impact of the performance difference between the volunteer computer on the entire group of task completed, and indirectly enhanced system responsiveness and reduce project completion time.
|
|---|
