The Research On Performance Tuning Of Mobile Database

With the further development of computer and wireless network technology, demands of applications on mobile devices such as mobile business are required increasingly. Meanwhile, when the quantity of data becomes larger with complexity of data processing higher, a lot of applications on mobile devices need embedded mobile database for data managing, thus, mobile database system emerges as the times require.Mobile database should be managed by special database managing system. Since the resource of both hardware and software in mobile devices is relatively limited than that of desktop computers, mobile DBMS should manage to optimize the system performance on the basis of full usage of current resources, to increase the throughput and response time of the database. On the other hand, because most users are not very familiar with the principles of databases, such performance tuning solution should be implemented automatically by the database itself, which is expected transparent to users so that the system can reach"zero management".SwiftDB is the prototype of a mobile databse management system developed by our team, it is applied in PDA and other mobile devices. In this paper, we first analyze current situation of adaptive management of database system, then based on the characteristics of both mobile computing environment and mobile devices, we propose the performance tuning policies applicable on SwiftDB: on the one hand, to save the memory space and increase the efficiency of applications implementation, mobile DBMS can automatically fits its functions reasonably into different applications, which is the customization solution for mobile systems, chapter 3 in detail introduces the policy for function customization in SwiftDB; on the other hand, to reduce the expensive I/O cost, mobile DBMS should automatically configure its system parameters suitable to current running environment, which is adaptive configuration technology for mobile databased, chapter 4 specifically discusses the policy for memory configuration and index selection.At the end of the pater, design for further development and complement is suggested for performance tuning of mobile databases.