OPTIMIZED FRAMEWORK FOR BANDWIDTH ADAPTIVE STREAMING OF LECTURE VIDEOS

Abstract

The explosive growth of the Internet has accelerated the development of multimedia applications. As video streaming applications are highly dynamic, due to vibrant nature of the audio/video contents. And also the network parameters are highly fluctuating in nature. So, the key challenges in deploying multimedia contents over the Internet are Quality of Service (QoS) requirements of multimedia applications and the demand on network resources. Our work presents a framework for E-Learning video streaming system to deliver scalable E-Learning video across the Internet to the clients. We exercised a very recent end-to-end efficient and scalable system for delivery of E-Learning videos. We complement the framework and provided a robust and accurate network bandwidth estimator for it. In-order to accurately estimate bandwidth we model the network algorithmically into two halves. The first part handles the prediction of network condition based on long-term network fluctuations. This part of the system is sensitive to the conditions that prevail in the network for longer durations. The second part handles the prediction of network condition based on short-term fluctuations. Our system tries to minimize the re-transmissions in vibrant networks. For this purpose we incorporated the noble EODist (Even — Odd distribution) concept to reduce number of re-transmissions and at the same time try to compensate for the lost frames. EODist is based on the fact of "the fundamentals of probability in communication theory" that the probability of receiving two consecutive erroneous frames is less than the probability of receiving single erroneous frame. The idea of EODist is to cluster even frames in one packet and odd frames in another packet for transmission; so at the receiver end, even if single packet of an even-odd pair is received error-free the de-compressor fills the gap of missing frames by performing interpolation on available frames of video. iii