Global Load Balancing Strategy Based on Origin-based Algorithm

Abstract

At present, most of the global load balancing products which undertake the resource access scheduling tasks for multiple data centers are mostly based on the single user optimal policy, they can’t reach global optimization, and in theory, algorithms like genetic and BP are mostly used, but they have slow convergence rate. The origin-based algorithm which is used for optimal planning of the transportation network is valued. To apply this algorithm, the model of resource access for multiple data centers which has corresponding attributes to the transportation network has been structured. It is proved that the model can be used in the actual global load scheduling scenario as under the same accuracy, the simulation result shows it can achieve a faster convergence, which is very important to meet the real-time scheduling requirements of multi data centers