Font Size: a A A

The Research On Autonomic Database Middleware

Posted on:2011-01-03Degree:MasterType:Thesis
Country:ChinaCandidate:Z X ZhaoFull Text:PDF
GTID:2178360305971738Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
With the increase in complexity of database management system and rise of DBA's cost, we urgent need database system with autonomic computing power to reduce maintenance costs and achieve the automatic configuration management of system according to different types of workload request. Therefore, the research of database system with autonomic computing features is paid more and more attention. In the meantime, the database system needs to deal with changing workloads in a network environment. In order to make the database system maintain a good state at the normal arrival of the workload or overloading, we need to monitor and forecast system operation status and workload classification status, and take measures to optimize the scheduling. According to system workload conditions, database middleware should be able to allocate resources rationally in the external database. If the workload consumes more resource, middleware allocates more resources to it. If the workload consumes fewer resources, the middleware allocates fewer resources to it to avoid waste of resources. This paper attempts to introduce certain autonomic computing features into the database middleware, which has some intelligence.The main work in this paper is that, the idea of autonomic computing and autonomic computing conceptual model are applied to the design of the autonomic database. Firstly, we put forward the concept and structure of autonomic database unit ADU and study an autonomic database system architecture, then present a state-based universal distributed autonomic database system architecture. Secondly, we build a database workload autonomic middleware model which consists of several components: workload listener, knowledge base, workload characteristic detector, system performance detector, performance predictor, performance evaluation, workload control scheduler, and give out the functions of various components.In the experiment, we use TPC-C transaction type as simulation workloads, adopt SQL SERVER 2000 database as the database system platform, introduce database autonomic middleware into the SQL SERVER 2000 database. The experimental results of database system performance operation verify the feasibility and effectiveness of the use of database middleware by comparing before and after the use of middleware.
Keywords/Search Tags:autonomic computing, autonomic database, middleware, workload scheduling
PDF Full Text Request
Related items