An Approach To Maximize Profits For Virtual Machine Allocation In Cloud Datacenters

In the years with the rapid expansion of the cloud computing, it has become the inevitable trend of the times that deposit the computing resource and the storage resource on the cloud. Currently, with the rapid increasing categories of the cloud and the quick improvement of the cloud quality, service providers are putting much energy into the relative technologies of the cloud service attempting to improve the profits of the cloud service.Researching the internal task scheduling mechanism of the cloud data center indicates that the scheduling system usually arranges multiple tasks to share the computing resource via one physical server to reach the goals of cost saving and usage rate improving. The shared scheduling way puts multiple tasks’ requests together which may affect the users’ response time. So it will be a very meaningful research about how to optimize the service profits and at the same time give consideration to the scheduling strategy performance.For raising the profits optimization technology which can give consideration to the data center with high efficient performance, the thesis first researches the profits optimization technology in current operational research and computer network science, analyzes their advantages and disadvantages and whether are suitable for the special scheduling environment as cloud service; then perfects the problem model of the interval scheduling and puts forward the capacity shared scheduling model theory which better fits the cloud environment. On the basis of this theory, raise a profits optimization scheduling method which applies to the data center scheduling system of the cloud computing. According to the analysis of the requirement parameter of the user virtual machine, can get the requirements of each virtual machine’s sequence arrangements in limited physics servers and hope for maximizing the cloud suppliers to provide profits of the virtual machines.Through the theoretic analysis, actual data test and contrasting with current main profits optimization technology, prove the arithmetic in this thesis having the effect to optimize the cloud service profits on the basis of the performance ensurance.