A CELLULAR NEURAL NETWORKS BASED DIFFSERV SWITCH FOR SATELLITE COMMUNICATION SYSTEMS

摘要

Recent developments of Internet services and advanced compression methods has revived interest on IP based multimediarnsatellite communication systems. However a main problem arising here is to guarantee specific Quality of Service (QoS)rnconstraints in order to have good performance for each traffic class.rnAmong various QoS approach used in Internet, recently the DiffServ technique has became the most promising solution,rnmainly for its simplicity with respect to different alternatives. Moreover, in satellite communication systems,rnDiffServ policy computational capabilities are placed at the edge points (end-to-end philosophy); this is very importantrnfor a network constituted by one satellite link because it allows to reduce the implementation complexity of the satelliternon-board equipments.rnThe satellite switch under considerationmakes use of theMultiple InputQueuing approach. Packets arrived at a switchrninput are stored in a shared buffer but they are logically ordered in individual queues, one for each possible output link.rnAccording to the DiffServ policy, within a same logical queue, packets are reordered in individual sub-queues accordingrnto the priority. A suitable implementation of the DiffServ policy based on a Cellular Neural Network (CNN) is proposedrnin the paper in order to achieve QoS requirements.rnThe CNNs are a set of linear and nonlinear circuits connected among them that allow parallel and asynchronousrncomputation. CNNs are a class of neural networks similar to Hopfield Neural Networks (HNN), but more flexible andrnsuitable for solving the output contention problem, inherent of switching systems, for VLSI implementation.rnIn this paper a CNN has been designed in order to maximize a cost functional, related to the on-board switch throughputrnand QoS constraints. The initial state for each neural cell is obtained looking at the presence of at least one packetrnfrom a certain input logical queue to a specific output line. The input value for each neural cell is a function of priorityrnand length of each input logical queue. The versatility of neural network make feasible to take the best decision for thernpacket to be delivered to each output satellite beam, in order to meet specific QoS constraints. Numerical results forrnCNN approach highlights that Neural network convergence within a time slot is guaranteed, and an optimal, or at leastrnnear-optimal, solution in terms of cost function is achieved.rnThe proposed system is based on the IETF (Internet Engineering Task Force) recommendations; this means that trafficrnentering the switching fabric could be marked as Expedited Forward (EF) or Assured Forward (AF), otherwise handledrnas Best Effort (BE). Two Assured Forward classes with different emission priority have been implemented, taking intornaccount time spent inside the logical queue and its length. Expedited Forward traffic is typical of services to be deliveredrnwith the maximum priority, as streaming or interactive services. The packets, belonging to services that need a certainrnlevel of priority with low packet loss, aremarked as Assured Forward. Best Effort traffic is related to e-mail or file transfer,rnor other that have not particular QoS requirements.