Virtual Network Mapping Through Locality-aware Topological Potential and Influence Node Ranking

Efficiently mapping multiple independent Virtual networks (VNs) over a common infrastructure substrate is a challenging problem on cloud computing platforms and large-scale future Internet testbeds. Inspired by the idea of data fields, we apply a topological potential function to node ranking and propose an algorithm called Locality-aware node topological potential ranking (LNTPR), which can precisely and efficiently reflect the relative importance of nodes. Using LNTPR and the concept of locality awareness, we develop the Locality-aware influence choosing node (LICN) algorithm based on a node influence model that considers the mutual influence between a mapped node and its candidate mapping nodes. LNTPR and LICN improve the integration of node and link mapping. Simulation results demonstrate that the proposed algorithms exhibit good performance in determining revenue, acceptance ratio, and revenue/cost ratio.