XML has become the de-facto standard currently for data representation and exchange onthe Internet, it already play an essential role in many fields. Meanwhile, it is interoperabilitystandards for Web services, document processing system and database. However, the queryprocessing of XML data has become a bottleneck that affecting the performance of thesesystems. Therefore, the demand of efficiently and fast XML processing is growing rapidly. Onthe hardware side, multi-core processors are increasingly becoming main stream on all kindsof computers. However, most XML processing algorithms still use the traditional serial mode,and thus can not make full use of the advantage of multi-core resources.In this paper, we use data partition for the XML document, in order to achieve loadbalancing, each thread process the sub-regional data independently in order to reducesynchronization and communication overhead between them. In data partitions, the maininquiry, under the three strategies based on document size balance and the balance based onthe number of nodes, in order to avoid the problem of load imbalance caused by idle on thelatter basis, the design strategy of a mission to seize, to avoid the CPU idle. By the queryprocessing time compared with single-core, three strategies can be effective to save querytime; compared the two data equalization method, the improved method can obtain part of theperformance advantages, thereby reducing network load and improve the efficiency ofdatabase queries and responsiveness. In these types of strategies through the analysis ofexperimental results, the results show that the policy is effectiveness and feasibility whenapplied to the Native XML Database System. |