3-LEVEL TRAFFIC SHAPER AND ITS APPLICATION TO SOURCE CLOCK FREQUENCY RECOVERY FOR VER VIDEO SERVICES IN ATM NETWORKS

In this paper we study the traffic shaping issues for VER video servic es in ATM networks with stress put on the source clock frequency recov ery aspect, We propose ''three-level traffic shaper'' (TLTS) which is able to limit the minimum as well as the maximum cell rate of input VE R traffic, For a VER video source modeled after a two-state MMPP, we d etermine, as performance measures, the cell loss probability, the unde rflow probability, and the maximum delay that a cell can experience in the proposed TLTS, Then we apply the TLTS for source clock frequency recovery in VER video services. In extending the existing CBR-based sy nchronous residual time stamp (SPITS) method for VER video services, i t is necessary to restrict the minimum, as well as the maximum, number of cells generated in a fixed period of time to ensure the transmissi on of timing information, and the proposed TLTS is tailored to meet th is necessity by adopting an optimal dimensioning procedure based on th e genetic algorithm (GA), Finally we provide some numerical examples t o demonstrate the optimal dimensioning procedure of TLTS and the relat ed performances of statistical multiplexing and source clock frequency recovery.