Framework for adaptive video streamingIn satellite networks

Abstract

Video streaming over satellite networks has to be tactfully handled because of it's inherent characteristic of high latency that leads to inefficiency if feedback mechanismlike TCP is used. Further if the satellite communication network includesdynamically shareable channels among different users, it really gets difficult todetect the bandwidth of each of the channel and accordingly manage the videobeing streamed through each of the channels.The main thesis of this research work is to estimate the bandwidth in such dynamicallyvarying satellite communication channels. This is attempted by usingthe tool of Pathload in satellite communication network. But due to issues of latencyand buffers, bandwidth is inaccurately estimated through this tool. Hence,a solution for these problems is proposed which results in accurate bandwidthestimation in the satellite networks. As part of the solution, the waiting period inthe Pathload code is adjusted as per the requirements of satellite communicationnetwork; additionally, the buffers present in the network path are not letting thebandwidth to be estimated properly. To solve this issue, the buffers are filled withsome dummy data before starting the actual measurement which led to accurateresults. Thus, this proposed logic is verified in this thesis by conducting numerousexperiments which were done both in the case of empty link and also in thepresence of cross-traffic in the link. Consequently, Pathload is customised withabove changes to estimate the bandwidth appropriately in satellite communicationnetworks.