Reserarch About The Layered Queueing Network Model Of The Database System In The Cloud Environment

As backstage of applications, database system is the core component of generating, supplying and manageing enterprise data. Database performance is in relation to the rising and falling of enterprise. The effective management for performance of the database system is an important task to ensure application performance. However, in the widespread popularity of cloud computing environments, because of cloud computing involves frequent task scheduling, complex data stream processing in the data processing process and other reasons, the performance of the database system under cloud environment is no longer as easy as traditional environment. In addition, due to limited resources, the cloud computing service providers always want to use the least hardware resources to achieve the highest customer satisfaction, so it is particularly important to predict the performance of database system in cloud environment.The core of predicting performance of the database system effectively is to build a database system performance model. A model that accurately describe the performance of database system can reduce the cost for the database system, and has an important role on ensuring its performance. For predicting the performance of the database system under cloud environment, based on structure of cloud computing and process of database systems analysis, the resources were stratified according to ideological layered queuing network model and made model. This model uses solutions of layered queueing network----MOL (the of Layer Method)algorithms and SRVN (the Stochastic Rendezvous of Network) algorithm to solve the model and to predict the performance of database systems. It shows Advantage of layered queuing network model in the database modeling system. Layered queuing network model of database systems has important implications for the capacity planning, resource allocation and performance management of database system.In this paper, firstly, the current studies on the method of predicting performance and the performance of cloud computing are interpretated and descripted. and then layered queuing network model of database system in cloud environment is proposed after a detailed introduction of layered queuing network model, combined with resource allocation of cloud computing. And this model is solved by solutions of layered queuing network model----MOL (the Method of Layer) algorithms and (the Stochastic Rendezvous of Network) algorithm. It validate the validity of the model. The main contribution of this paper is:first, layered queuing network model of database system is proposed with shared-memory situations under cloud environment. It describes the behavior of the database system in the way the model; second, the SRVN (the Stochastic Rendezvous of Network) algorithm is achieved. and MOL (the Method of Layer) algorithm and SRVN (the Stochastic Rendezvous of Network) algorithm are contrasted in the algorithm structure, model limitations, the ability of describtion that facilitate the selection for the two algorithms in the future.In this paper, the DB2database is as an experimental tool, TPC-H is for the loading.This paper base on layered queueing network model. After a lot of repeating experiments, this model obtain an average error rate of10%-20%.This paper verifies the effectiveness of layered queuing network model of database systems in the traditional database system and in cloud computing environment.It has important significance on performance of database system.