Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
451800 | Computer Networks | 2014 | 20 Pages |
Cloud computing is a promising paradigm which has emerged to overcome the main issues of the computational world. It acts as a torchbearer technology for realising a new computing model in which resources can be acquired and released on demand. However, a fundamental issue in the instantiation of resources is how to afford optimal allocation so that the service provider fulfils the users’ service level agreement while minimising its operational cost and maximising its revenue. In this paper, we tackle the problem of networking static service provisioning within the Cloud’s private backbone network. This requires the embedding of virtual networks in which edge routers are directly connected to data centres. Our objective is to map online virtual networks in the private substrate backbone network using the minimum physical resources but while still satisfying the required QoS in terms of bandwidth, processing power and memory. This in turn minimises the reject rate of requests and maximises returns for the substrate network provider. Since the virtual network embedding problem is NP-hard, we propound a new scalable embedding strategy named VNE-AC to deal with its computational hardness. This is based on the Ant Colony metaheuristic. Extensive simulations are used to evaluate the performances of our proposal. These show that VNE-AC minimises the reject rate of virtual networks and enhances the cloud provider’s revenue.