Research And Implementation Of Workload Control Algorithm In Workload Adaptation DBMS

With rapid development of science and technology, information technology has been treated as a symbol of productive forces, information data also grow rapidly. Especially with the development of network technology, vast amounts of network data need to be store and manage in database. So, the network database comes out in this time. The traditional application of database management methods which rely on the DBA, only can adjust the parameters to meet the requirement, however, it can not meet the new practical requirement.Workload adaptation database technology is performance optimization technique. It uses a new method to manage the database system. The cooperation of Canada Queen's University and IBM Toronto Lab proposed a novel framework for workload adaptation databases system. In this framework, people can improve the performance of storage system by monitoring the different factors. This paper proposed a middleware model of workload adaptation database system based on this framework, paying great attention to the research on workload control model. The key work of paper can be summarized as follows:Firstly, Proposed a middleware model of workload adaptation database system,the model consisted by five components, they are interceptor, workload characterization module, system monitoring module, repository and workload control module. Interceptor used to intercept the workloads; workload characterization module used to predict the characteristic parameters of the workload; System monitoring module used to monitor the execution information of the workload and the resource usage of the system; repository used to hold the characteristic attribute of the workload and the resource usage of the system; workload control module used to control and filter the workload.Secondly, in the workload adaptation database, how to control and filter the workload is a typical multi-objective optimization problem. No-dominated Sorting Genetic Algorithm (NSGA-Ⅱ) is a multi-objective optimization algorithm, it used to solve the multi-objective optimization problem. In this paper, based on the analysis and research of the NSGA-Ⅱtheory, we introduced a new workload control algorithm which using the NSGA-Ⅱ. Using NSGA-Ⅱalgorithm select some workloads which have bigger commercial value, shorter response time and consume less resource will first be submitted to the database. According to the factors that affect database performance, build a multi-objective optimization model for workload control, which is the basis of workload control.Finally, Using NSGA-Ⅱalgorithm to control the workload, there may be some individuals which they are not satisfy the constraint conditions during the process of genetic operation, this will affect convergence speed. We will amend the individuals by introduced greedy algorithm into NSGA-Ⅱ, which can improve convergence speed of NSGA-Ⅱalgorithm. We verified the improved NSGA-Ⅱby experiment,it can improve convergence speed in some extent.