Quality of Service Implementation for Transmission of Video Data - Phase 2

Principal Investigator:

Vladimir Cherkassky, Professor, Computer Science and Engineering

Project Summary:

This project is investigating the practical transmission of multiple video streams over limited-bandwidth communication channels. This includes transmission over limited-bandwidth wireless channels, as well as transmission over fixed-link channels where the bandwidth requirements exceed available link capacity.

Previous research (Phase I) focused on the technical issues important for quality of service (QOS), such as video compression (using off-the-shelf commercial hardware/software) and prioritization of video data in packet-switching (IP) networks. These issues have been addressed using the prototype system developed/installed at the ITS Lab at the University of Minnesota. Phase I research helped to understand and quantify the trade-offs between the amount of compression, quality of video, available network bandwidth, and varying network traffic loads. Phase II research is concerned with practical design choices for implementing QOS under ?typical? network configurations used at Mn/DOT. This includes implementing the QOS prioritization software in the prototype system at the ITS Lab; analyzing existing network configurations and application settings that require QOS implementation; and simulating these network configurations in the prototype system in order to evaluate practical effectiveness of the QOS approach for different application settings. In Phase II, potential bottlenecks that might disrupt data flow at a time of congestion in the Mn/DOT networks were identified. Also, a prototype system for studying the effects of QOS, configured to provide video streams from a set of video servers, has been set up in the ITS Lab. An open source QOS software has been configured as part of the prototype system to provide Quality of Service to the video streams. In the future, different real-time scenarios will be simulated in the prototype system and the behavior of the QOS software under these settings will be extensively studied.


Project Details: