In this work, we propose a priority buffer management scheme for MPEG-2 video traffic called MPFD. The MPFD algorithm is based on constructing a virtual buffer that represents the future buffer occupancy using information about video frames within a future lookahead period. The virtual buffer is then used to determine whether incoming low priority frames need to be dropped in order to protect future high priority frames. MPFD is compared to existing buffer management schemes that has been used for video traffic. MPFD shows a significant frame loss reduction as compared to other buffer management schemes.

In this work, we present an algorithm to analytically estimate the queueing performance of MPEG-2 video using goodput as the metric; where we define goodput as the ratio between the number of cells in uncorrupted and correctly displayable frames to the total number of cells that arrive at the queue. The estimation algorithm is used to evaluate three buffer management schemes. The effect of congestion at the output link is also investigated. The algorithm produces a good approximation of the frame goodput metric and closely agrees with the corresponding simulation results.

We present an algorithm to estimate analytically the frame goodput of a correlated MPEG-2 video source that traverses a first network node's buffer; where frame goodput is defined as the ratio of the number of cells in completely and correctly transmitted frames through the buffer to the total number of cells arriving at the buffer. The effect on the frame goodput of changing the output-to-input line speeds is also investigated. By using this algorithm, a buffer that is three times the maximum frame size is seen to yield loss probability of 10/sup -6/ at a normalized output line rate of 0.28 for a load of 20%. A maximum simulation error margin of /spl plusmn/30% for values of the order of 10/sup -5/ with a 98% confidence was achieved.

We consider broadcast WDM networks operating with schedules that mask the transceiver tuning latency. We develop and analyze a queueing model of the network in order to obtain the queue-length distribution and the packet loss probability at the transmitting and receiving side of the nodes. The analysis is carried out assuming finite buffer sizes, non-uniform destination probabilities and two-state MMBP traffic sources; the latter naturally capture the notion of burstiness and correlation, two important characteristics of traffic in high-speed networks. We present results which establish that the performance of the network is a complex function of a number of system parameters, including the load balancing and scheduling algorithms, the number of available channels, and the buffer capacity. We also show that the behavior of the network in terms of packet loss probability as these parameters are varied cannot be predicted without an accurate analysis. Our work makes it possible to study the interactions among the system parameters, and to predict, explain and fine tune the performance of the network.