On the physical and logical topology design of large-scale optical networks

We consider the problem of designing a network of optical cross-connects (OXCs) to provide end-to-end lightpath services to large numbers of label switched routers (LSRs). We present a set of heuristic algorithms to address the combined problem of physical topology design (i.e., determine the number of OXCs required and the fiber links among them) and logical topology design (i.e., determine the routing and wavelength assignment for the lightpaths among the LSRs). Unlike previous studies which were limited to small topologies with a handful of nodes and a few tens of lightpaths, we have applied our algorithms to networks with hundreds or thousands of LSRs and with a number of lightpaths that is an order of magnitude larger than the number of LSRs. In order to characterize the performance of our algorithms, we have developed lower bounds which can be computed efficiently. We present numerical results for up to 1000 LSRs and for a wide range of system parameters such as the number of wavelengths per fiber, the number of transceivers per LSR, and the number of ports per OXC. The results indicate that it is possible to build large-scale optical networks with rich connectivity in a cost-effective manner, using relatively few but properly dimensioned OXCs.     

Call admission control schemes: a review

Over the last few years, a substantial number of call admission control (CAC) schemes have been proposed for ATM networks. We review the salient features of some of these algorithms. Also, we quantitatively compare the performance of three of these schemes     

Open Networks of Queues with Blocking: Split and Merge Configurations:

Split and merge configurations of open networks of queues with blocking are analyzed approximately. The approximation procedure developed decomposes these queue networks into individual queues with revised queue capacity and revised arrival and service processes. These individual queues are then analyzed in isolation. The results obtained seem to have an acceptable error level. Stability conditions for these two configurations are also obtained.     

Self‐Propelling and Positioning of Droplets Using Continuous Topography Gradient Surface

A radial pattern with continuous topography gradient is presented, which induces a continuous inward wettability gradient and enables self‐propelling and accurate positioning of droplets to the pattern center. The effect of droplet size and wettability gradient of the pattern on the self‐mobility of droplets is investigated. The wettability gradient is found to increase towards the pattern center, enhancing the self‐motion of droplets at the inner area of the pattern. Moreover, larger droplets give rise to a larger solid‐liquid contact diameter, which helps to satisfy the self‐motion criteria that the advancing contact angle at front edge is smaller than the receding contact angle at rear edge. Consequently, a larger droplet size favors self‐motion initiated from the outer area of the pattern. The continuous topography gradient employed here allows the flexible dispensing of droplets at any place within a certain range, and avoids potential pinning defects to droplets at geometrical discontinuities. An average self‐motion velocity up to 4.0 cm/s for microliter‐sized droplets is achieved on the resultant patterned surface.     

Semaphore queues: modeling multilayered window flow controlmechanisms

An open queuing network for analyzing multilayered window flow control mechanisms consisting of different subnetworks is presented. The number of customers in each subnetwork is controlled by a semaphore queue. The queuing network is analyzed approximately using decomposition and aggregation. The approximation was validated against exact numerical and simulation data, and it was found to have an acceptable relative error. The queuing model is easy to implement, and it can be included in a software package. Using this queuing network, a case study involving the modeling and analysis of the ISO X25 flow control mechanism is presented     

Optical burst switching for the next generation Internet

Demands for network bandwidth increase daily. In order to meet our ever-increasing network bandwidth needs, one solution is to design and build the next generation Internet with an optical core infrastructure, with last connection provisioning time and unprecedented high data rates of 100 terabits per second and higher. An optical network is built by interconnecting various optical switches with wavelength-division multiplexing (WDM) fibers, i.e., fibers that can simultaneously transmit data over different wavelengths. Many of today's commercial optical networks, however, do not utilize the WDM technology efficiently. With respect to the current state of the technology, an Optical Burst Switched (OBS) Network is one of the most promising all-optical architectures for the next generation Internet. It efficiently supports the transmission of bursty traffic over an all-optical infrastructure. OBS is still being developed and it has not been standardized yet. This article describes the main features of an OBS network, its benefits as well as its challenges.     

A discrete-time queueing model of the shared buffer ATM switch with bursty arrivals

We model the shared buffer ATM switch as a discrete-time queueing system. The arrival process to each port of the ATM switch is assumed to be bursty and it is modelled by an interrupted Bernoulli process. The discrete-time queueing system is analyzed approximately. It is first decomposed into subsystems, and then each subsystem is analyzed separately. The results from the subsystems are combined together through an iterative scheme. The analysis of each subsystem involves the construction of the superposition of all the arrival processes to the switch. Comparisons with simulation data showed that the approximate results have a good accuracy.Supported in part by DARPA under Grant No. DAEA18-90-C-0039.Work done while on a sabbatical leave of absence at the Computer Science Department of North Carolina State University.     

Scheduling Combined Unicast and Multicast Traffic in Broadcast WDM Networks

This paper studies the performance of various strategies for scheduling a combined load of unicast and multicast traffic in a broadcast WDM network. The performance measure of interest is schedule length, which directly affects both aggregate network throughput and average packet delay. Three different scheduling strategies are presented, namely: separate scheduling of unicast and multicast traffic, treating multicast traffic as a number of unicast messages, and treating unicast traffic as multicasts of size one. A lower bound on the schedule length for each strategy is first obtained. Subsequently, the strategies are compared against each other using extensive simulation experiments in order to establish the regions of operation, in terms of a number of relevant system parameters, for which each strategy performs best. Our main conclusions are as follows. Multicast traffic can be treated as unicast traffic, by replicating all multicast packets, under very limited circumstances. On the other hand, treating unicast traffic as a special case of multicast traffic with a group of size 1, produces short schedules in most cases. Alternatively, scheduling and transmitting each traffic component separately is also a good choice.     

A Comparison of Allocation Policies in Wavelength Routing Networks*

We consider wavelength routing networks with and without wavelength converters, and several wavelength allocation policies. Through numerical and simulation results we obtain upper and lower bounds on the blocking probabilities for two wavelength allocation policies that are most likely to be used in practice, namely, most-used and first-fit allocation. These bounds are the blocking probabilities obtained by the random wavelength allocation policy with either no converters or with converters at all nodes of the network. Furthermore, we demonstrate that using the most-used or first-fit policies gives an improvement on call blocking probabilities that is equivalent to employing converters at a number of nodes in a network with the random allocation policy. These results have been obtained for a wide range of loads for both single-path and general mesh topology networks. The main conclusion of our work is that the gains obtained by employing specialized and expensive hardware (namely, wavelength converters) can be realized cost-effectively by making more intelligent choices in software (namely, the wavelength allocation policy).