The Research On Data Sharing Platform Distributed Data Access And Load Balancing

In the data sharing platform, in order to integrate the old system, enable data exchange between different departments, with the XML package heterogeneous data formats, different systems and different environment with a uniform data format, distributed data Transparent access. UDDI registry by service providers for the services up, the service requester through the UDDI registry registered service interface, service providers can transparently access the services. Business data exchange platform with more and more requests for more tasks, the need for UDDI to make it to transform the function of scheduling, task allocation to achieve the load, and because a larger amount of data for data exchange, the network transmission Will appear in an intolerable delay, so that it can slice parallel XML data.Thesis, according to data exchange platform, using the Web Service and JMS messaging middleware technology for distributed data access, the use of the service to publish Web Service using the SOAP protocol for transmission to XML format for data exchange, using the J2EE platform, using JMS Message middleware request/release distributed access. Application of data exchange platform, the client requests more frequently, no previous scheduling system, there services to respond to the obstruction, the dimensionless UDDI scheduling this paper, the task of scheduling, the use of fair scheduling algorithm. Used in data transmission line based on the query data partitioning algorithm.This paper introduces the data exchange platform for the domestic research, discusses the status of data exchange platform for data exchange platform features, Web Service technology as the basis for selection, the use of JMS messaging middleware, the use of XML for packaging, in-depth analysis, etc. Introduction.Secondly, the main technology for distributed data access business systems design and implementation. Distributed data access, mainly relying on JMS message queue, to achieve the release of service, the request asynchronously. Messaging middleware rely on a message key mechanism for data sharing platform transparency visit.Finally, this system has been modified so that it has a slice-based task scheduling and load balancing. Based on the obtained load information, to borrow a fair scheduling algorithm, and the defenders of a task queue, and the same task at the request of the optimization method proposed. For the return of large data blocks, in accordance with the check out method, the fragmentation, it has a parallel transmission capability, this model was optimized, and the transformation of the system before and after the transformation of the simulation test.