Cloud Computing Environment Load Balancing Mechanism

Cloud computing is based on software services, such as virtualization and distributed computing. It is the direction of the next generation of network development. Differed from cluster or other systems, there are some important features, just like delivered service and usage patterns. Cloud computing relies on some computer technology. It contains not only grid computing and distributed computing, but also load balancing. Therefore, how to achieve load balancing becomes one of the crucial issues of cloud computing. Under the cloud computing environment, load balancing mechanism needs to take into account with its effectiveness, robustness, and scalability. Based on its characteristic, the paper proposed three load balancing mechanism:the load balancing mechanism based on the improved FCM algorithm, honeybee-based and improved honeybee-based. It is more effective in load balancing mechanism based on the improved FCM algorithm. The mechanism which puts the original FCM algorithm and the PSO algorithm together classified the load status of the server compute nodes accurately. And on this basis, part of the job on high-load server emigration. Due to reasonable migration, this system has a greater improvement in effectiveness than others'. Honeybee-based load balancing mechanism performance is most excellent in robustness. It analogy user request(severice) to each bee. Each bee will carry nectar information by itself. After nectaring, the information will bring back to the hive. The rest of the bees determine the nectar position through the analysis of this information. The entire process is to seek the largest quantity of nectar. As each round of nectar have scouts to join, so the whole optimization process is not easy to fall into local optimum. As a result, the throughput of the system will not be affected significantly with the more or less of the request. Last, improved honey bee mechanism has been proposed with better scalability. Similar to PSO algorithm-based load balancing mechanism, the improved one added a judgement which limited the waiting queue length. It prevented request from lineing in the little contacted server queue. This initiative reduced the unnecessary waiting for these request. The main job is to:(1) To classify the type of request and use different load balancing strategies to different types of requests, makes the cloud computing management node serve professional and better. Under the premise to ensure load balancing effect, as far as possible to improve its processing speed.(2) Combined Fuzzy clustering with load balancing mechanism. In order to put the computing nodes togeter into consideration, computing nodes are classified to5classes. Instead of separately address each node, it speed up the processing speed of the system.(3) Improved the correctness of classification in FCM algorithm.(4) Combined the honey bee algorithm with load balancing mechanism, and improved honey bee algorithm to enhance the scalability of the mechanism.