Prediction Algorithm For Load Balancing Mechanism In The Cloud

Cloud computing is a new type of information service environment which is established on the basis of the Internet in recent years. It uses existing data center technology, Internet technology and core technology such as intelligent terminal technology to produce a kind of efficient, convenient service which allow users to use the existing resources in the cloud computing environment. At the same time, all kinds of service resources paid per use. C loud computing environment, the supply of service resources to make people's lives more convenient, but the process of providing services is not simple. In cloud computing environment, user? demand for services in the stages of change in time, when the service resource increase or decrease. How to realize the management and configuration of service resources in the cloud computing environment has a very important influence on the service quality of the cloud computing environment.In this paper, we study how to effectively avoid the unreasonable allocation of service resources and the unbalanced load of service resources when faced with a large amount of service demand. How to put an end to the phenomenon of partial service resources idle, part of the overload of service resources, improve the overall service capabilities of cloud computing environment. In view of the above problems, this paper mainly does the following aspects of research.First, do the selection and processing of data. Select LLNL Thunder? cleaned data from ?Logs of Real Parallel Workloads from Production Systems Logs?. Select the desired attributes, data trimming process, the data into the required format.Then, compare and select the suitable prediction algorithm. The main comparison and application is in smoothing index prediction model, time series linear prediction model and artificial neural network prediction model. By using historical data to predict the next stage of data, and to compare the forecast results with the real data, the forecast indexes of each method are obtained. Comparative analysis of various forecasting methods and then carried out, a way to obtain high accuracy.Finally, select the forecasting method to predict real data analysis, simulation by cloudsim, obtained experimental results. The results demonstrated that load balancing based on the prediction is effective.