Introduces a very effective priority mechanism for high speed DQDB type metropolitan area networks (MANs). The proposed mechanism can be applied to either of the two variations of the no slot wasting bandwidth balancing mechanism and enables high priority traffic to have almost immediate access to the transmission medium; regardless of the presence of low priority traffic. The authors use simulation results to investigate the delay performance of the proposed priority mechanism in the case of three classes of traffic. Furthermore, they compare its performance with the corresponding performance of other (previously proposed) priority mechanisms. Finally, they examine its ability to support real-time traffic such as voice and video.>

The no slot wasting immediately using the TAR bit (NSW IUT) bandwidth balancing mechanism has been previously proposed for dual bus architectures. NSW IUT is a variation of the no slot wasting bandwidth balancing (NSW BWB) mechanism and it can also introduce bandwidth fairness into a DQDB network without wasting channel slots. In addition, it enables lightly loaded stations to use more effectively the TAR=1 bits they observe and improve considerably their delay performance. In the paper, the authors use simulation to carry out a thorough investigation of the delay performance of the NSW IUT mechanism under single and multiple priority classes of traffic as well as under the presence of erasure nodes. Throughout the study, they also compare NSW IUT with NSW BWB as well as with the bandwidth balancing of the IEEE 802.6 standard.>

The no-slot-wasting bandwidth balancing (NSW BWB) recently proposed for dual-bus architectures provides throughput fairness and arbitrary bandwidth distribution without wasting channel slots. The performance of two variations of NSW BWB in the presence of erasure nodes and under one or multiple priority classes of traffic is studied by simulation. For both variations, the effect of the erasure mode locations on the throughputs of the various stations, as well as priority classes of traffic, is examined, and the performance of NSW BWB is compared with the corresponding performance of the bandwidth-balancing mechanism of the distributed queue dual bus. The simulation results reveal some very interesting properties for the first of the two NSW BWB variations that are used to derive analytic estimates of its throughput performance in the general case of arbitrary number of stations and arbitrary location of erasure nodes.>

A thorough investigation of the performance of the rotating slot generator (RSG) scheme, based on simulation, is presented. RSG is a medium access control protocol appropriate for high-capacity long-distance metropolitan area networks (MANs). It uses the looped bus architecture of the distributed queue dual bus (DQDB) in which the slot generators for both busses are colocated inside the same station. However in RSG, all the stations, one after the other in a cyclic order, undertake the task of generating and destroying the slots on both busses. In this way the location of the station relative to the slot generator changes dynamically, and its effect on the performance is drastically reduced. The authors investigate the fairness and performance of RSG under symmetric and asymmetric loading, underload and overload conditions, and under the presence of a single or multiple priority classes of traffic. They also compare its performance with different variations of DQDB.>

A throughput and delay analysis of FDDI and FDDI-II is presented under mixed traffic. The case of voice/data transmission is considered, and the throughput and delay behavior of FDDI, under packetized voice transmission, and FDDI-II, under isochronous voice transmission, is examined. The performance of the two networks is then investigated under the transmission of voice, data, and video. Two transmission alternatives are considered. In the first, all three types of information are transmitted in packets over an FDDI network. In the second, the isochronous slots of FDDI-II carry the voice and video. The effect of various system parameters on the performance of both networks is examined, and their voice, video, and data performance are compared.>

A novel approach to the delay analysis of various service disciplines in symmetric token ring networks is presented. It is shown that exact or approximate expressions for the average packet delay of different service disciplines can be directly derived by the delay expression of the exhaustive service system. This is accomplished by: (a) inflating, by an appropriate factor, the packet size of the discipline whose performance is analyzed; and (b) using the inflated packet in the delay expression of the exhaustive service system. The proposed method is applied to the limited service and time token systems, and its accuracy is verified by simulation results. Advantages of the method are its higher accuracy over previous analytic methods and its robustness to ring latency and other system parameters.>

The authors present performance results for medium access control protocols which are appropriate for voice/data integration in high-bandwidth token-passing networks. They consider two classes of protocols: (1) multiple priority cycle protocols and (2) timed-token protocols. To provide bounded delays for voice packets, both types of protocols use timers to control the data-packet transmissions on to the channel. The effect of the number of active voice sources on the timers' values is discussed. Simulation results are presented indicating that for both protocols the most important performance parameter is the maximum data information that can be carried by the channel, when the number of active voice sources is specified. Analytic formulae are derived which accurately predict the data capacities for both protocols. The derived formulae are used to present the effect of several system parameters on performance as well as to compare the performance of the two types of protocols.>