LAUSR

Current video content providers require a huge demand for network resources to obtain the best quality of experience from the end-users perspective. This measure can be achieved if the network can anticipate itself to different existing issues (bandwidth problems, etc.). Due to traditional Internet Protocol networks cannot handle this problem due to the rigidity of its architecture, software defined networks is introduced as a potential answer for resolving the existing challenges in video content delivery decoupling the control plane and data plane. Utilizing the unified perspective of the network, it is possible to develop a monitoring tool obtaining the network metrics to be used afterwards. In this paper, we propose the application of different clustering algorithms for monitoring software defined networks to validate the number of statistics queries decrements using the data and error rates to improve the network traffic and to reduce the overload. The simulations demonstrate that the expectation-maximization algorithm highly reduces the monitoring queries maintaining the best data and error rate values in the video delivery.