The Representation And Organization Of Knowledge Rules In Workload Adaptation Database

With the explosion of portfolio, the prolongation of burgeoning businesses, the popularization of virtualization technology and the pursuance of office productivity, to mitigate the pressure of database server and ensure high availability, enterprise data center must chose the new equipment with high property, warranted service level and simple management, so that it can defend the continuity of business, improve the quality of service and support the development of new business. However, it means huge investments and more database managers were needed, which would improve the running cost of enterprise tangibly virtually; different from the traditional hand-operation, database workload adaptation actualizes the optimization of database system performance automatically and independently through the reasonable control load of executive sequences and opportunities based on the differences of the utilization of system resources and load values under the same software and hardware resources conditions, so that it can promote the utilization of system resources and reduce the costs of manpower.Contra pose the researches nowadays, in this paper, we carried out on the knowledge base component that based on the models of WAAD (Workload Adaptation Architecture for DBMS).In the WAAD, knowledge base component is the important component in system, which stores results of system-monitoring component, workload control component and priority of workload by DBA and so on. The knowledge base component finishes analyzing, processing the data and carries out necessarily deduction. With the help of knowledge base component, we reduce the costs of the computing scale of workload control.This paper mainly introduces as follows:In this paper we study the related of the Workload Adaptation Architecture with knowledge base for DBMS on the basis of improved WAAD. Methods of knowledge acquisition and expression are proposed according to the database load adaptive system's characteristics and working process.In order to judge whether there is rules in knowledge base can be used quickly, Bloom Filter is introduced. The paper present a new kind of Bloom Filter-DSIBF (Dynamic Segment Incremental Bloom Filter), which tears the data structures into several small segments and adds the same size segment to represent a set. Then we analyze it with the Bloom Filter in detail in probability of false positives and accommodate elements. We prove that DSIBF can efficiently solve the problem of dynamic growth element data set and this method can reduce the overall waste of storage space.At last, corresponding test platform for knowledge base is built to evaluate the effect of use knowledge base and did not use knowledge base.