Adaptive Load Balancing Algorithm. Campus Network

Load balancing algorithm in load balance technology of cluster is an important question in the research of campus network application and the whole network application. Aiming at characteristic of requirement changing rapidly in enterprise management software, in this paper, we introduce, Application Integration Technique based on campuse network is becoming the mainstream in the development of Education administration Information management system. However, in present, between education softwares lack of interactive, and the level of application integration is low, those reason lead to sharing and communicating information in campuse difficult. To sovle this problem, we use the conception and technology of enterprise application integration to solve the problem of the information isolated island. Analyzing campuse network, the paper introduces a new method of cluster nod load, and discusses two load balancing algorithms base on queuing theory and based on culture algorithm, in the end, I wrote the program of Clustersimulator to realize those algorithm.In the research of interactive framework in education information management system, we found as the number of the servers in framework increasing and the scale of system largening and the data of commucation excessing. There is necessary to use load balance technology of cluster to strengthen processing proformance in server cluster.In this paper, research the main load balance algorithm, because of the defects of represent method in load; I propose a new method of represent in load. It not only considers the hard index load, but also considers the soft index load, the continuity and forseebility jugement of input queue. And validate the improvement of this represent method.After that, this paper introduces two load balance algorithms based on queuing theory and based on cluture algorithm.In the load balance algorithm based on queuing theory, according to the present load situation and the rest processing ability, compute the need average arrival frequency of server in fixed time intervals, and then get the task allocation strategy. Also a simulative experiment system was developed.In the load balance algorithm based on cluture algorithm, I fully using the the thought of cluture evolutionary, apply into load balance algorithm. The allocation strategy of this algorithm is that population evolution adopt adaptive crossover and mutation operator, because of active in long time and mutiply, some excellent cluture information is saved in population and some bad information is discard. The cluture is improved continuously, and the population is evolution continuously. After some generation, there will be some excellent individual. In this paper, there is description and designing in details of algorithm base on cluture algorithm, and a simulative experiment system was developed. We validate the improvement in the load balance algorithm.