The study of algorithms for Multi-VMs bandwidth guarantee in datacenters

With the development of the server virtualization technology and the rapid growth of network applications, datacenters have received significant attention because of its cost-effective and promising for large date storage, HPC, cloud computing and diversified network services and applications. Similar to the CPU and memory resource, the network should be shared by co-located tenants in a manner of proportionally to their payments, and meanwhile the minimum bandwidth guarantee for VMs and the utilization of network is also should be taken into consideration. In this paper, on the basis of existing researches, we introduce the algorithms of Multi-VMs bandwidth guarantee. The preliminary simulation results show that we can deploy more VMs with a small probability of congestion. If the VM sends large amounts of data, the algorithm in the paper would have a better performance in network utilization and transmission time.