Network Stability Based on the Amount of Available Bandwidth in a Software Defined Networking

Abstract

The modern day Network operators, several network users and some sensitive applications running in the network environment have recognized the impact of the network available bandwidth (ABW) estimation as significant due to its importance. The significance is very unique most especially to video streaming applications, peer-to-peer video conference and the selection of some improved dedicated services from network servers. Moreover, determining the amount of available bandwidth within the network can help in designing a guaranteed network stability mechanism within a typical scalable network. Several algorithms and tools were proposed in the literature to estimate the amount of bandwidth available within a network, but the application of the estimation has been unused to address optimal network performance in relation to network stability. Therefore, this paper utilized the improved calibrated pathload algorithm that is enhanced by the correlation coefficient technique to improve the network stability. This ABW information is exploited to direct critical flows over a suitable path through the installation of OpenFlow rules on the switches. These rules were installed likewise as modules into the controller. The initial performance evaluation showed that our proposed approach significantly improved both the bandwidth utilization and the stability of the network in comparison to the ticket probing approach.