Performance-Aware Optimal Controller Placements Via Genetic Algorithms For Software-Defined Networking

Abstract

Software-Defined Network Is A New Paradigm Of Computer Networking That Is Based On The Concept Of Separating The Control Plane And Data Plane To Simplify Network Management Through Central Network Programmability And Rapid Innovation. Multi-Controllers Designing Is A Promising Way To Achieve Reliability And Scalability. However, It Brings The New Problem Of Controller Placement In A Distributed Architecture. Non-Optimal Controller Placement Will Reduce The Performance By Increasing Delay, And Decreasing The Throughput Of Controllers. The Existing Controller Placement Based On Simulated Annealing, And Controller Placement Based On Particle Swarm Optimization Are Two Approaches That Are Recently Developed To Solve Such Problems In Which CPPSO Is More Optimal Than CPSA. But Still Both Existing Approaches Are Not Effectiveoptimal Placements, Which Reduces The Performance And Scalability Of The Systems. To Solve These Problems Controller Placement Based On A Genetic Algorithm Has Been Proposed In Our Study. The Proposed CPGA Used The Fitness Value Of Each Node To Locate The Controllers At Their Optimal Place. Also, Our Study Used The GA Operations Until It Gets The Optimal Placement Of Controllers And After Locating The Controller At Their Appropriate Place, It Was Used For A Long Time In The Case Of Near Optimality Rather Than The Existing Approaches. The Implementation Was Simulated Using The MATLAB Simulation Tool. The Performance Of The Proposed Solution Has Been Evaluated By Using Throughput And Propagation Delay Metrics And Compared With CPSA And CPPSO. Accordingly, The Simulation Results Indicated That CPGA Improves Throughput By 5.2% Over CPSA And 2.6% Over CPPSO, And Improves Propagation Delay By 12.4% Over CPSA And 5.7% Over CPPSO. This Study Concludes That The CPGA Indicates A Better Improvement In All Assessed Metrics That Shows Efficient Results Over CPSA And CPPSO By Locating Controllers At Their Optimal Place