Performance Evaluation of Sign-Language Videoconference Traffic

This paper presents modeling results for sign language videoconference traces (H.261 and H.263 encoded) which were captured from realistic multipoint IP videoconference sessions between Greek signers. By comparing traffic models against simulations, it is proven that the application of the proposed theoretical parameters into a generic continuous Markovian model can lead to a conservative, though accurate, solution for the analytical treatment of this type of traffic