A New Algorithm for Virtual Networks Reconfiguration with Adaptive Interval
Qing Lin, Yulei Huang, Yan Li · 2018
Reconfiguration can reduce the bottleneck nodes/links when virtual network has been embedded successfully, with virtual network requests coming and expiring along the time. However, well-designed virtual network embedding algorithm considers load balance or other constrains to map virtual nodes/links, and the virtual nodes/links would not be mapped on bottleneck nodes/links. So, there are two extremes in substrate network: no bottleneck nodes/links or almost all nodes/links are bottleneck nodes/links; the reconfiguration algorithm based on substrate network would not work or virtual nodes/links would be migrated repeatedly. In order to solve these problems, we provide virtual networks reconfiguration algorithm with adaptive interval, in which we trigger reconfiguration action based on evaluating virtual network record to describe average resource utilization ratio of substrate nodes/links which bears virtual nodes/links, and calculates adaptive interval between reconfiguration actions. Simulation results show that the provided algorithm can get more smooth time intervals and reconfiguration cost, higher acceptance ratio and resource utilization ratio than the compared algorithms when substrate network can bear appropriate virtual networks.