Research On The Delay Constrained Virtual Software Defined Network Embedding Problem

Software Defined Networking(SDN)is believed to be a promising networking practice in facing the challenges of ever-increasing traffic demand and network growth.With wider deployment of SDN enabled network infrastructures,the conventional network virtualization problem will gradually evolve into an SDN virtualization problem,in which both virtual data plane and control plane entities are instantiated from their physical network counterparts.The instantiation of the data plane and the control plane are called the data plane mapping and the control plane placement respectively.Furthermore,in wide-area network(WAN),owners of vSDN may restrict propagation delay between SDN controller and virtual nodes to satisfy a variety of demands;it would influence both the data plane mapping and the control plane placement.In this thesis,we first argue that the virtual SDN embedding problem is in general different from the conventional virtual network(VN)embedding problem.We try to define the vSDN embedding problem space,followed by a potential network architecture that supports SDN virtualization.We then formulate the vSDN embedding problem with the objective of maximizing the long-term request acceptance ratio and propose a heuristic to tackle the embedding problem.Our results show that both the data plane mappinging and the control plane placement are restricted by a strict controller-to-node delay constraint,and there is a mutual constraint between the data plane mapping and the control plane placement.These would have a major impact on the vSDN request acceptance ratio.Beyond that,the diversity of vSDN request would also have a certain extent effect on the request acceptance ratio.