Research And Design Of Cloud Data Center Network And Dynamic Bandwidth Scheduling Technology Based On SDN

Currently,technologies,such as mobile Internet,the Internet of Things,and cloud computing,are burgeoning.To cope with the problems arising in efficient storage and data processing,the construction of cloud data center based on large-scaled distributed computing is now in the ascendant globally.Software-defined network(SDN)offers a feasible solution for these problems properly.It achieves the decoupling on control and data layers and provides a practical solution for the realization of network virtualization.Additionally,it plays an increasingly important role in network architecture design since it simplifies network architecture design and enhances the flexibility of network deployment.Under the background of the SDN technology-based network construction of cloud data center,investigations were given to the bandwidth scheduling and energy consumption optimization of cloud data center.It was attempted of realizing the flexible scheduling and management to the network resources in cloud data center and optimizing resource utilization efficiency.1)The shortcomings of the delay scheduling in traditional data center were firstly analyzed.Then based on SDN's network architecture and Hadoop's system framework,an improved Hadoop delay scheduling algorithm,known as BS-IDS,was introduced.This algorithm realized the effective control on data transmission owing to its effective adaptability to the dynamic load changes of data center.Subsequently,the performances of FIFO,traditional delay scheduling algorithm,and BS-IDS,were tested.Test results verified that the latter presented superior performance and efficiency in terms of the adaptability to the load changes of data center,the distribution efficiency of task,and the time cost of data migration.2)An energy consumption optimization plan was proposed for cloud data center based on the migration and integration of virtual machines.Regarding server's energy conservation,a multi-objective optimization model was established to solve the migration and integration problems of virtual machine.In addition,hybrid partheno-genetic algorithm(HPGA)was employed to more rapidly obtain the optimal solution;in reference to network's energy conservation,a network energy conservation control mechanism was put forward based on flow sensing.Using the topological and flow sensing abilities of SDN controller,the energy conservation topology was computed to control the sleeping and awakening of network devices,so as to realize network's energy conservation.