Research And Implementation Of Load Balancing Based On Predicting Under Cloud Computing Environment

In recent years, cloud computing technology has made widely used in the industry, lots of companies migrate their site system platform to cloud. Along with this process, the user confidence is fast-growing on cloud platform, the scale of cloud hosting sites is expanding rapidly. Cloud computing technology make computing,storage and networking hardware resources as a resource pool, provide users on-demand computing resources. However, under current technical conditions, the size of a single instance of a user who can get can not bigger than a single physical server. To solve this problem, cloud platform add load balancing technology.Traditionally, load balancing technology mainly used in computing clusters, load balancing algorithms currently used are mainly designed for the compute cluster.Therefore, how to improve the load-balancing technology to make them take full advantage of the characteristics of cloud platforms and improve resource utilization, is a subject that worth study.After analyzing the theoretical principles of the current cloud computing and load balancing technology, based on the features of current applications that deployed on the cloud platform,we propose a load balancing scheduling algorithm based on load prediction. The algorithm considers the fact that the applications deployed on cloud platform are website, base the analysis of real data on users visiting of a web site,then compare a variety of prediction algorithms, we choose the BP neural network as the prediction model of this load balancing algorithm. Scheduling part of the algorithm has considered the feature that the users can easily apply and release resources on cloud platform, based on the load forecast results, the scale of the resources owned by the user that can be dynamically adjust, this will improve resource utilization.OpenStack is an open source cloud platform that has gain the biggest part of cloud market, we choose OpenStack as the base platform for achieving the load balancing algorithm. This paper analyzes the OpenStack composition and architecture,then using its existing load balancing component, choose python,and shell as programming language with cron, curl and other open source software to achieve its implementation.In this paper,we build an OpenStack cloud platform and then deploy the load balancer, we use the real world data to simulate. The result shows that: after using this load balancing algorithm, users can effectively reduce idle computing resources.