The primary focus of this project is to monitor service level quality in networks where the service in question is independent of the network provider(s) and is a streaming application possibly with interactivity leading to strict quality and latency requirements (e.g., voice/video regular or conference calls). The scales and the variety/diversity of such networks and services are ever increasing, thus generating large amounts of specific traffic. The project aims at analyzing and optimizing such traffic.
Videoconferencing is a convenient way for meeting with people while avoiding transportation costs. With the advent of cameras on every laptop and smartphone, it has become a commodity feature delivered by well-known products such as Skype, FaceTime and Hangouts. However, live video streams have latency requirements and consume bandwidth. Furthermore, videoconferencing is not efficiently implemented in current applications as it often uses a central server and does not rely on network layer services. With the advent of Software Defined Networking, it is now possible to use advanced techniques such as multicasting and stream layering inside the network for optimizing live video transmission. The following code was written to investigate the impact of these techniques on the bandwidth and latency of Scalable Video Coding videoconferencing calls as well as the call capacity of such software defined networks.
The source code is released under the GNU LGPL version 3.
Christelle Al Hasrouty
Last updated: April 26, 2018.