Infrastructure Deployment And Optimization Of Integrated MicroDC-NGPON Networks For Green Cloud Computing

This thesis integrates a new Micro Data Centers (MicroDCs) with Next-Generation Passive Optical Network (NGPON) for the purpose of achieving the advantages of cloud-computing services, such as good expansibility, strong mobility, simple deployment, localization services, low costs, etc. Compared with conventional large data center, i.e. Meg-DC, this integrated network architecture has higher economic benefits, lower transmission delay, less energy consumption and better expansibility. In this thesis, we research mainly on the deployment and optimization issues of MicroDC-NGPON integrated network. Aiming at minimizing the deployment costs of the integrated network, this thesis can not only calculate the number of network components, but also determine the best positions of network components. This thesis targets at providing schemes for completing the planning and deployment of "last mile cloud service" MicroDC-NGPON network.We first establish mathematical models for IDO issue, then use Gurobi to solve this model. And the result demonstrates that the IDO issue proposed by us is solvable. Although we manage to gain the optimal deployment results, the solving time of Gurobi experiences an exponential growth. Aiming at the problem of high complexity, we suggest a new heuristic algorithm to solve the IDO issue for the sake of obtaining the optimal results and improving the efficiency. In the simulation experiment, we conduct the deployments for different network to verify the effectiveness of the network models. Through the comparison of the two methods, we can conclude that the new heuristic algorithm is suitable for large-scale network deployment and enables us to have the optimal results quickly.