Fast packet switches for variable-size packets have become an everyday necessity with the rapid growth in the volume of Internet traffic. Such switches can be designed in two different ways, either by segmenting packets into smaller fixed-size cells and designing packet switches for such cells, or by designing generic packet switches for variable-size packets, where packet segmentation and reassembly can be omitted. The second option is investigated in this paper. The synchronous operation mode with time-limited bulk service is selected. The switching fabric is assumed to be internally non-blocking and provided with input queues. A previous maximum switch throughput analysis has been done under the assumption that the length of the time slot is fixed set to its minimum allowed value (Tmin). In this work, a so-called time-slot stretch factor (SF) is introduced. The actual time-slot length is determined by multiplying Tmin with the SF, where SF. Next, a so-called Internet traffic-source model is proposed based on findings from real IP traffic measurements. The performance implications of the proposed time-slot length modification are analyzed by discrete-event computer simulation. The maximum switch throughput is increased by increasing the SF value, e.g. for uniform packet size distribution and SF=10, the maximum switch throughput is increased from 75% to 97%. The influence of the traffic-source characteristics on the maximum switch throughput is decreased when SF value is increased. In order to prevent any possible throughput degradations, it is advisable to use integer SF values. Packet delay analysis has revealed that by increasing the SF value, the mean packet delay is also increased. Nevertheless, it is shown that the number of switch input and output ports is the most important factor to be considered when packet delay is at stake. Service class differentiation inside investigated packet switch is possible and is not affected by the increasing SF value. Such a packet switch is suitable for implementation in wide area networks, due to high transmission speeds and the small number of switch ports.

Real traffic flows are captured in various network environments and their statistical properties are analyzed. Based on real traffic flows, MWM (Multifractal Wavelet Model) and Poisson equivalent synthetic traffic flows are generated. Performance analysis of a SB (Synchronous Bulk) packet switch is joined with different types of traffic. Maximum throughput performance of the SB packet switch for various real traffic flows and appropriate MWM and Poisson equivalent synthetic traffic flows are evaluated by using discrete-event simulations. Different flow persistence, SF (Stretch Factor) and scheduling mechanisms are used in order to asses their influence on SB packet switch performance. Traffic asymmetry, either input or output based, has a major influence on SB packet switch performance. By increasing the level of asymmetry, maximum throughput values decrease considerably, especially if the ROT (Rotation) scheduling mechanism is applied. Traffic asymmetry also decreases the influence of the SF parameter on maximum switch throughput. As a general rule of thumb, SF values of no more then 5 must be used if asymmetrical traffic is switched. It is also advisable that OPF (Oldest Packet First) scheduling mechanism is used in such cases. The influence of burstiness and scaling of traffic flows turns out to be relatively insignificant for the SB packet switch maximum throughput results, if the OPF scheduling mechanism is used. Larger throughput discrepancies are detected, if ROT scheduling is used.