A Resource Node Selecting Algorithm Based On Network Distance In Regional Cloud

The development of cloud computing has brought a new revolution to the Internet industry, but also brought a lot of opportunities and challenges. Region-oriented cloud computing platform is growing rapidly, while users using cloud services continue to increase, how to quickly schedule cloud resource between region-oriented cloud has become an important research scheme in order to meet the requirements of the quality of cloud service. The present study on cloud resource scheduling are mostly for proprietary network resource scheduling, which assumes that this proprietary network is accessible network in which the bandwidth resource is infinite and propagation delay is zero, while for the region cloud which is across the WAN, this assumption is not valid. For this reason, when it comes to the region cloud resource node between data centers, we must consider network conditions and select the resource nodes of the network in good condition to ensure rapid resource scheduling during inter-regional cloud node.in view of WAN characteristics of region-oriented cloud, this paper presents regional cloud resource scheduling model based on P2 P technology, based on the model, the author has analyzed the networking situation between the requesting nodes and serving nodes, focusing the impact of the bandwidth, packet loss rate and delay on the scheduling efficiency, determined the Odds Rate over this three factors with AHP according to the experiment to calculate the weight. As a result, the author has proposed nodes selecting algorithm based on network distance while resource scheduling, which has judged the result of the factors values multiplied by their weight as the distance valve, according to which, the resourcing nodes with good performance are selected to schedule the resource.In this paper, in view of WAN cloud characteristics of region-oriented cloud, with the use of the technology of P2 P resource scheduling, the author has analyzed the networking situation between the requesting node network nodes and serving nodes, focusing the impact of the bandwidth, packet loss rate and delay on the scheduling efficiency, has proposed a scheduling model, based on this model named "network distance", the requesting nodes can choose the optimal serving nodes to obtain the resource.Under the simulated network environment of LAN and WAN, compared with the random opting nodes algorithm and the algorithm based on big bandwidth, the algorithm proposed in this paper has a good performance.