A Load Balancing Protocol for Improved Video on Demand in SDN-Based Clouds

Abstract

Video streaming has emerged to become an essential component of communication methods on the internet. However, streaming of video still faces a number of challenges. The transmission of video content over the Internet entails stringent bandwidth, delay and jitter requirements to guarantee better visual quality. The Internet, however, only offers best effort service that does not guarantee any requirement in terms of delay, loss and bandwidth, making video streaming very complicated to achieve. Existing routing protocols do not always offer assurance that certain quality of service will be met. In order to improve video streaming, it was proposed to transform the cloud architecture and its operations by incorporating the software-defined network approach. This also suggests a rethink in the design of routing protocols for video streaming. Several software-defined networks-based routing algorithms are proposed, however they do not always address the viewer's demand for a certain quality. This research introduces the Load Balancing Routing Protocol (LBRP) as an adaptive routing protocol that improves user-experience of stored video streaming applications. LBRP accomplishes routing by considering parameters such as link capacity and re-routing non-real-time traffic towards under utilized links. The performance of LBRP was tested using Mininet and taking into consideration metrics such as bandwidth distribution, throughput and delay. The results showed that LBRP gives precedence to real time traffic over non-real-time traffic in the network providing improved viewing experience for the user.