Design, Implementation and Evaluation of Virtual Resource Description and Clustering Framework

This paper presents an approach to speed up and enhance matching of virtual network requests to available resources in virtual network provisioning frameworks. The method consists of introducing a weight or score expressing the importance of the resources, their attributes and the values taken by these attributes. The scores are obtained through statistical analysis of the requests for virtual network resources. Previous methods treat resources equally and experience longer delays and lower efficiency during the matching process. They were originally not designed for virtual network provisioning and the proposed method in this work aims at addressing their weaknesses in scalability, in resource representation and incremental updates. To highlight this new method, a functional comparison with these previous techniques is proposed. Evaluation results of the score based clustering and matching algorithm show improved scalability and performance in clustering efficiency and matching delays for virtual network provisioning.