We propose an efficient heuristic algorithm that sets up and releases lightpaths for connection requests dynamically. We partition the routing and wavelength assignment (commonly known as RWA) problem into two subproblems and solves both of them using a well-known shortest path routing algorithm. For solving the routing subproblem, an auxiliary graph is created whereby the nodes and links in the original network are transformed to the edges and vertices, respectively, and the availability of each wavelength on the input and output links of a node as well as the number of available wavelength converters are taken into account in determining the weights of edges. Furthermore, for solving the wavelength assignment subproblem, an auxiliary graph is also utilized and the cost for wavelength conversion is taken into consideration in the edge weight function. A distinguished feature of our algorithm is that it employs more accurate network information on the availability of both the wavelengths and the wavelength converters than the existing algorithms in deciding the routing and the wavelength assignment. Simulation results show that our algorithm performs much better than previously proposed algorithms with comparable computation time, especially when the number of wavelengths is large while the number of converters at each node is limited.

In the basic user/network interface of N-ISDN (ITU-T Recommendation I.430), the D channel is shared by up to 8 terminals for signal and data packets. An analytical model is proposed to reveal the performance characteristics of the access control scheme for the D channel. Numerical and simulation results are shown to demonstrate the performance differentiation of the terminals with different priorities. It is observed that the mean signal delay at a low load become large because of the long service time for packets and that the priority mechanism may not work properly when the loads at the terminals are very asymmetric.

The commenters identify and correct errors in some of the equations in the above-titled paper by Ferguson and Aminetzah (see ibid., vol.COM-33, p.223-31, Mar. 1985), particularly those related to gated service discipline. In addition, they compare equivalent gated and exhaustive systems (using the corrected equations) numerically in terms of the mean waiting time behaviors at the various nodes, and in terms of the overall mean and variance of the mean waiting times across the nodes.>

A two-layer performance model of a local-area network (LAN), consisting of a media-access control (MAC) layer submodel and transport layer submodels, is presented. A multiple-queue cyclic-service model of the token-passing MAC layer is used along with a closed-queueing network model for the transport layer. To deal with acknowledgement traffic with priority over data messages, a mean-value analysis (MVA) priority approximation is devised for the case where the priority can change at each queue. An iterative solution algorithm is proposed for this two-layer performance model. The method is applied to five models of communication systems: a symmetric-load, piggybacked acknowledgement model: a symmetric-load, explicit-acknowledgement model; full-duplex communication models (without and with priorities); and a client/server model. System performances measures such as throughput and mean-message delay, computed from the analysis, are compared to the simulation results over a wide range of parameters. In most numerical examples, the approximate analysis is shown to be within 90% confidence intervals in the mean-message delay, and within a few percent error in the throughput.>

In token passing protocols of local area networks, we assume that a user transmits a message which is divided into several packets, and that each packet is transmitted in a round of the circulating token. This transmission from a certain station is interleaved with possible transmissions from other stations. In this paper, a new queueing model is provided to evaluate the mean message delay in contrast to the mean packet delay given by the past studies. The same model is also used to evaluate the mean packet delay in error-prone transmission systems.

Analysis is given to a nonpreemptive priority queueing system with P classes of messages where the class of message to be served next is the highest priority class waiting at the time of service start. (If this were the highest priority class waiting at the service completion epoch, we would have a classical nonpreemptive head-of-line priority queueing system.) We assume that the message service time distribution is identical for all classes. The mean message waiting time is obtained explicitly for each class, and numerically compared to the values in the corresponding head-of-line system. We have also proposed and evaluated a fairness measure to demonstrate the degree of discrimination. This model can be applied to the performance analysis of the prioritized token-ring scheme in local area computer networks when the propagation delay and bit latency are negligible compared to the frame transmission time.

The processes consisting of the packet interdeparture times for contention-type packet broadcasting systems in a hidden-user, singlehop environment are studied under the heavy-traffic assumption. The channel access protocols considered include pure ALOHA and unslotted nonpersistent carrier-sense-multiple-access (CSMA). The theory of superposition of independent renewal processes is applied to approximate the distribution of the duration of each unsuccessful transmission period in channel state. Our analysis results for the channel throughput and the coefficient of variation for the packet interdeparture time in symmetric configurations are shown to be in good agreement with simulation results over a wide range of offered channel traffic.

The channel throughput for a finite number of packet broadcasting users is analyzed for random access protocols, including slotted persistent carrier sense multiple access (CSMA) with and without collision detection and unslotted persistent CSMA with and without collision detection. We consider bothp- and 1-persistent CSMA. Our results can be extended to infinite population cases (by taking the proper limit), where they agree with the known throughput expressions when available.

We consider a system of two users of slotted CSMA-CD (carrier-sense multiple-access with collision detection). The two users are assumed to have independent identical packet arrival streams, the identical randomizing policy for retransmission, and an infinite capacity for storing queued packets. The mean packet delay (including the queueing and retransmission delays) is derived explicitly.

The processes consisting of the packet interdeparture times in contention-type packet broadcasting systems are studied under the heavy-traffic assumption. The channel access protocols considered include slotted and unslotted ALOHA and carrier-sense-multiple-access (CSMA) with and without collision detection. Through analysis of the Channel activity cycle, the distribution, mean, and coefficient of variation of the packet interdeparture times are explicitly derived. Taking the reciprocal of the mean interdeparture time, we obtain the channel throughput. Cases with dissimilar users are mainly considered, and systems of statistically identical users are treated as special cases.

A Markovian model is formulated to find the throughputdelay performance for slotted-ALOHA multihop packet radio networks with a fixed configuration of packet radio units (terminals and repeaters) and fixed source-to-link paths for packets. Improvements in performance which are obtained by the adjustment of transmission parameters (suppression/acceleration) according to the states of nearby units and/or by having repeaters equipped with multiple buffers are demonstrated.

In multihop packet radio networks with randomly distributed terminals, the optimal transmission radii to maximize the expected progress of packets in desired directions are determined with a variety of transmission protocols and network configurations. It is shown that the FM capture phenomenon with slotted ALOHA greatly improves the expected progress over the system without capture due to the more limited area of possibly interfering terminals around the receiver. The (mini)slotted nonpersistent carrier-sense-multiple-access (CSMA) only slightly outperforms ALOHA, unlike the single-hop case (where a large improvement is available), because of a large area of "hidden" terminals and the long vulnerable period generated by them. As an example of an inhomogeneous terminal distribution, the effect of a gap in an otherwise randomly distributed terminal population on the expected progress of packets crossing the gap is considered. In this case, the disadvantage of using a large transmission radius is demonstrated.