Font Size: a A A

Research On Volunteer Computing System Architecture And Configuration Management

Posted on:2012-03-04Degree:MasterType:Thesis
Country:ChinaCandidate:X W WangFull Text:PDF
GTID:2248330395458121Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
With the popularization of computers and Internet, increasingly more scientific projects take advantage of Volunteer Computing System (VCS) to converge and then employ mass of computing resources spreading over the whole Internet, which makes research of VCS gradually become an important research field.Nodes’anonymity and high dynamicity bring a host of technical problems to the design of VCS, such as resource management, fault-tolerance, security etc. As the foundation of other system function, configuration management is the core content of the design of VCS. It needs not only to solve problems of startup and initialization of VCS, but also to solve description and management of resources needed by system runtime. As a result, configuration management is one of the most difficult problems for VCS design.At first, this paper summarized technical features of volunteer computing and components and characteristics of existing VCS. Based on the analysis of design purpose and functional components of VCS, this paper proposed a VCS named PPVCS, which is based on JAVA language, P2P technology and node group.Then, this paper designed PPVCS’s function components in detail. For starter, this paper designed two manners of system start-up, process of quitting PPVCS and strategy of node management. In node management, this paper improved SCAMP algorithm with trust mechanism; Furthermore, this paper combined distributed task pool and task’s lifecycle to manage tasks. Finally, this paper detailedly described different protocols and message formats used in different scenarios.At last, this paper implemented partial functions of configuration management. We implemented two manners of system start-up and initialization, processes of message passing and data structure used by message process methods used by nodes and tasks, defined the way to description nodes, based on which we implemented methods for managing node’s dynamic join and leave and failure.
Keywords/Search Tags:Volunteer Computing, System Architecture, Configuration Management, NodeManagement
PDF Full Text Request
Related items