ON THE PERFORMANCE OF EARLY PACKET DISCARD

In a previous paper, one of the authors gave a worst case analysis for the early packet discard (EPD) technique fur maintaining packet integ rity during overload in ATM switches. This analysis showed that to ens ure 100% goodput during overload under worst case conditions requires a buffer with enough storage for one maximum length packet from every active virtual circuit. This paper refines that analysis, using assump tions that are closer to what we expect to see in practice, and examin es how EPD performs when the buffer is not large enough to achieve 100 % goodput, We show that 100% goodput can be achieved with substantiall y smaller buffers than predicted by the worst Ease analysis, although the required buffer spate can be significant when the link speed is su bstantially higher than the rate of the individual virtual circuits. W e also show that high goodputs can be achieved with more modest buffer sizes, but that EPD exhibits anomalies with respect to buffer capacit y, in that there are situations in which increasing the amount of buff ering can cause the goodput to decrease. These results are validated b y comparison with simulation.