GSOCPP optimization for predicting the proper number of controllers in SDN

Document Type : Research Paper

Authors

1 Department of Computer Engineering, Central Tehran Branch, Islamic Azad University, Tehran, ‎Iran.‎

2 Department of Computer Engineering, Central Tehran Branch, Islamic Azad University, Tehran, ‎Iran.

3 Department of Computer Engineering and ‎IT,‎ Amirkabir University of Technology, Tehran, ‎Iran‎.

Abstract

In Software Defined Network (SDN), the controller layer that is separated from the data layer is responsible for all system functionalities. However, centralized solutions suffer from single-point-of-failure and bottleneck problems. Several controllers are used to increase availability and performance in large networks to solve the aforementioned problems. One of the main concerns is finding the optimal number of controllers and their locations, which is known as an NP-hard problem. To do this, in this paper, in addition to presenting an efficient algorithm based on Garter snake algorithm (GSO), a new statistical analysis for determining the number of controllers is figured out.

Keywords