Video traffic is expected to become increasingly important withthe large scale deployment of broadband ISDN. In the literature, itsuggested that smoothing variable bit rate (VBR) video traffic beforetransmitting it onto the network would help reduce the probability ofpacket loss. The authors show why deterministic smoothing at the sourceapproximates the minimum achievable loss over the network end-to-end ofall possible smoothing schemes. Furthermore, they develop a powerful yetsimple analytical technique that can efficiently calculate the lossprobability at any point in a network carrying video traffic. Theyvalidate the analytical results using traces of actual video segments.The results can be used for admission control and traffic management.They find that in the case of highly correlated traffic such as video,the way to control loss is to ensure that the fraction of time thearrival process exceeds the service process is small
展开▼
