We propose two multipriority reservation protocols for wavelength division multiplexing (WDM) networks. The network architecture is a single-hop with control channel-based passive star topology. Each station is equipped with two pairs of laser and filter. One pair of laser and filter is always tuned to wavelength λ0 for control and the other pair of laser and filter can be tuned to any of data wavelengths, λ1, λ2, ..., λN. According to the access methods of the control channel, one protocol is called slotted ALOHA-based protocol and the other protocol is called TDM-based protocol. The two protocols have the following properties. First, each of them has its own priority control scheme which easily accommodates multipriority traffics. Second, they can be employed in the network with limited channels, i.e. the number of stations in the system is not restricted by the number of data channels. Third, they are conflict-free protocols. By using a reservation scheme and a distributed arbitration algorithm, channel collision and destination conflict can be avoided. For the performance point of view, the TDM-based protocol gives an optimal solution for the priority control. However it is less scalable than the slotted ALOHA-based protocol. The slotted ALOHA-based protocol also performs good priority control even though it is not an optimal solution. We analyze their performances using a discrete time Markov model and verify the results by simulation.

This paper investigates an ATM multiplexer with a resume level, which uses a selective cell discarding strategy as a priority control, and a Markov-modulated deterministic process (MMDP) as the burst input traffic. Assuming that a system is loaded with a superposition of several independent and homogeneous On-Off bursty sources with two priority classes, we obtain the cell loss probability of each priority class of an ATM multiplexer with a resume level. The performance analysis derived here includes as special cases one without priority and one with a threshold level. From the numerical results, we compare the cell loss probability, the mean queue length, the mean queuing delay, the level crossing rate, and the queue length distribution at the embedded points for the case of a threshold level with those for the case of a resume level. By selecting an appropriate resume level, we can reduce the sensitive state change around the threshold level.

Two simple and high performance multichannel distributed queue dual bus (MDQDB) protocols based on the wavelength division multiplexing (WDM) network technology are proposed, and the network architecture and operations are presented. To be suited for a high speed network, they inherit the advantages of the DQDB protocol such as node simplicity, network flexibility and distributed operations of individual nodes. The network capacity can also be greatly increased with marginal increase of node complexity. Simulation has been done to estimate the performances such as throughput and average access delays for individual nodes. The influence of the bandwidth balancing mechanism on the two protocols is considered at medium, high, and overload conditions. We also investigate the fairness characteristics in asymptotic conditions for various initial states.

In this work, teletraffic performance of a hierarchically overlaid microcell and macrocell structure in a PCS environment is presented. In this system, each group of N microcells is overlaid exclusively by one macrocell. The microcell tier is dedicated to low-mobility users and the macrocell tier is dedicated to both high-mobility users and overflowed low-mobility users from microcells. In general, forced termination is considerably undesirable from the user's viewpoint compared with the occurrence of call blocking. Therefore, handoff calls are given access priority to channels at the macrocell tier. As a priority scheme at the macrocell tier, a partial buffer sharing (PBS) scheme is proposed. As performance measures, call blocking probability, forced termination probability, and carried traffic are derived. Effects of the sizes of a macrocell and of hierarchical cell structure on system performance are discussed. The performance of the PBS scheme is also compared with those of the no priority scheme (NPS) and the first-come-first-service (FCFS) queueing scheme. Numerical results show that the PBS scheme gives the best performance for handoff traffic among the three schemes.

In this work, the performance evaluation of a voice and data integrated fast PRMA has been done. In this newly proposed protocol, every data terminal has an infinite buffer and voice terminals can work independently of data terminals. Thus voice terminals have the higher priority over data terminals. We can, therefore, analyze voice and data subsystems separately. For voice analysis, we use a Markov analysis, and we use the EPA method for data analysis in order to create an analytic form. As performance measures, the voice packet dropping probability, the average voice and data delays, and the total throughput have been derived. It is shown how many voice and data terminals the fast PRMA protocol can accommodate in a frame under the constraints that the voice packet dropping probability should be less than 0.01, and average data packet delay should be less than 250 msec. We also discuss the stable region for this system. Numerical results show close agreement between analysis and computer simulation.

A loop architecture of DQDB with slot reuse (LDQDB-SR) segmented by erasure nodes was studied to overcome the performance limitation due to the nature of the unidirectional bus architecture of DQDB with slot reuse. The LDQDB-SR adopts the destination slot release and an inter-segment bandwidth regulation based on the distributed queuing system of DQDB. This network suffers not only from severe throughput deterioration due to a high regulation cost and an excessive transit delay but also from unfairness in bandwidth sharing, especially under an overload condition. In this paper, we introduce an enhanced loop architecture of DQDB with slot reuse (ELDQDB-SR) to improve the performance of LDQDB-SR. The ELDQDB-SR uses a quota-based inter-segment bandwidth regulation mechanism to effectively control the bandwidth use of each segment. Each station selects the bus that minimizes the number of erasure nodes on the path to destination stations. Fairness control methods of DQDB are reviewed and the alpha-tuning mechanism is modified to achieve a fair bandwidth distribution among stations within each segment. Simulation results show that the ELDQDB-SR gives an enhanced throughput level and also maintains good fairness under overload conditions.