Font Size: a A A

Energy Saving And Load Balancing For SDN Based On Multi-Objective Particle Swarm Optimization

Posted on:2017-05-15Degree:MasterType:Thesis
Country:ChinaCandidate:R S ZhuFull Text:PDF
GTID:2308330485982530Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the rapid development of cloud computing and data centers, the problem of network energy consumption is increasingly prominent. Nowadays in the modern net-work path redundancy, unbalance link utilization and heavy energy consumption is ubiquitous problems exist.Most of the energy saving strategies on current IP network only aggregate traffic into a part of links. Though it can really save some power consumption,it leads to im-balance link utilization and seriously impacts the quality of service. How to achieve energy saving and load balancing under the premise of network transmission perfor-mance is the main problem to be solved of current network. With the emergence of the software defined network, the intelligent energy management makes solving this prob-lem possible. SDN is a kind of new network paradigm and it divides the network con-trol plane and data forwarding plane, what improves the network resource utilization and simplifies the network management. SDN allows the controller to realize central-ized control of the entire network by separating the network control plane and forward-ing plane.It has a prominent advantage on traffic control and power management and it provides a better platform for the implementation of the energy-saving strategies.In this paper, we take advantage of the centralized control and global vision of SDN network architecture to achieve the network energy saving and load balancing while ensuring QoS and transmission performance by designing energy aware routing optimization algorithm. The main work is as follows:(1)We first focus on how to maximize energy saving while ensuring load balancing. We take load balancing as constraint conditions and energy saving as the only optimiza-tion objective.Then a mixed integer programming model is established for this mul-ti-constrained maximum parallel flow problem. This problem is NP, so it is difficult to work out a feasible solutionin in a short period of time in large-scale networks using traditional algorithm. We introduce swarm intelligence to the routing optimization problem and design particle swarm algorithm is to solve it.(2)We futher study how to achieve the best optimization objective of load balanc-ing and energy saving at the same time, which means that we take load balancing and energy saving as our optimization at the same time.Then a multi-objective mixed inte-ger programming model is established. The multi-objective maximum parallel flow problem is NP-hard, so we improved the traditional particle swarm optimization algo-rithm and first propose multi-objective particle swarm optimization algorithm called MOPSO to solve this problem.(3)At last we do simulation experiments based on real topologies and traffic de-mands and results show the effectiveness of our algorithm both on the objective of en-ergy saving and load balancing compared with other algorithms.
Keywords/Search Tags:Software Defined Network, Multi-objective Particle Swarm Optimization, Energy saving, Load balancing, Mixed Integer Programmin
PDF Full Text Request
Related items