Research And Implementation Of Cloud-Based Virtual Network Management System

Cloud computing is a technological revolution of the IT industry, and has become the future direction of development of the IT industry. This change makes the degree of specialization in operating IT infrastructure continue to concentrate and improve. In the use of cloud computing, users lack the ability to configure the network, which is not open to the user. Although virtual network services of cloud computing has got more concern of cloud computing providers than before, their support in this regard is still in the imperfect stage.Based on cloud computing, this article designs a virtual network management system in order to provide the ability for users to configure virtual networks. A unified network technology architecture is established in this system based on the cloud computing environment, in order to provide users with network services based on their needs. This system is composed of Cloud Controller and Network Controller. Cloud Controller is responsible for accepting user deployment strategies, managing physical resources, virtual machine placement and allocating storage. Network Controller implements the communication between virtual machines and assigns IP addresses to virtual machines. Besides, Network Controller achieves security requirements of network traffic isolation by way of VLAN, and can optimize the deployment of user needs, by using a modified mapping algorithm.In the process of actual deployment of user needs, we improve and use a Weighted Reusable Mapping Algorithm to optimize the placement of virtual machines and their networks. This algorithm uses a two-step structure including node mapping and link mapping. Greedy algorithm is adopted in node mapping process, and k-shortest path algorithm is adopted in link mapping process. In node mapping process, the value of the remaining resources of neighboring nodes is weighted according to the principles of proximity, so that the neighboring nodes can get the priority of being selected. Besides, the set of candidate nodes is expanded by utilizing the reusable feature of nodes in the deployment of virtual machines, resulting in lower difficulty of link mapping. The advantage of this algorithm is that it takes into account the location of the virtual machine and virtual network constraints, so that resource utilization is improved. Simulation results show that this algorithm has a higher acceptance rate of user demand, higher earnings in the average revenue, and good performance in terms of resource utilization.