Design And Simulated Implementation Of Virtual Network Embedding Algorithm In Elastic Optical Data Center Network

With the rapid development of Internet, bandwidth demand is booming in a high rate. This not only makes network resource tend to saturation, but also leads to the variety of bandwidth requirements. Therefore, the construction of flexible and large capacity optical network becomes particularly important. In the traditional Wavelength Division Multiplexing (WDM) optical network, allocating network resource in a one-size-fits-all manner, leads to inefficient resource utilization and low flexibility. However, by introducing the Orthogonal Frequency Division Multiplexing (OFDM) technology and breaking through the fixed bandwidth spacing restriction between wavelength tunnels, elastic optical network utilizes spectrum resources efficiently, thus it has a wide application prospects. At the same time, with the widespread use of cloud computing and virtualization technology in data centers, Virtual Network Embedding (VNE) becomes an important challenge in the cloud data center network. It enables the co-existence of multiple virtual networks on the same substrate network by sharing the available resources. Therefore, virtual network embedding problem in elastic optical data center network has been widespreadly concerned. Meanwhile, server consolidation technology and spectrum defragmentation technology based on virtual concatenation are core technologies in the cloud data center network and elastic optical network respectively. How to apply them to virtual network embedding problem in elastic optical data center network becomes a new challenge.This thesis describes key technologies used in virtual network embedding problem, mainly including server consolidation technology, spectrum allocation constraints and spectrum defragmentation technology based on virtual concatenation. In addition, this thesis describes virtual network embedding problem in elastic optical data center network and the mathematical model of virtual network embedding problem is also presented. For static traffic, this thesis designs a virtual network embedding algorithm based on the layered auxiliary graph. And on this basis the thesis proposes an algorithm referred to as Virtual Network Embedding based on Virtual Concatenation (VNE-VC). This algorithm can consolidate the fragmented spectrum as continuous and available frequency slots for subsequent requests, improving the spectral efficiency for elastic optical network. In addition, this thesis proposes virtual network embedding algorithms based on four different ordering strategies according to the service order of requests.This thesis develops the simulation and analysis for the proposed algorithms. Simulation results show that, in terms of improving network resource utilization, the proposed virtual network embedding heuristic algorithm based on virtual concatenation reflects good performances. In addition, this algorithm can also reduce the blocking rate of virtual optical network requests. Therefore, it can be used as a valuable reference to build an elastic optical data center network.