Self-similar traffic is distinguished by positive correlation, which can be exploited for better traffic management. Inspired by measurement-based admission control schemes, a measurement-based congestion alarm is proposed. The aggregate traffic at an output port of a switch or router in a high-speed network is modeled by a fractional Gaussian noise process. Traffic measurements are performed in regular time intervals to determine the current traffic loading. This information is then used to predict the loading situation in the near future. If congestion is likely to occur, a congestion alarm is set off and appropriate network management functions taken to alleviate the possible congestion. The above constitutes a closed loop feedback control mechanism that maintains high resource utilization. Simulation results show that the proposed scheme, when used with dynamic bandwidth allocation, reduces bandwidth requirements by more than 20%.

In this paper, we study the use of reroutable assignment for multipoint videoconferences in a high-speed network. A conference model is constructed and conference calls are classified. A conference of a particular type can ride on different route-configurations. According to the location of the current speaker, a conference has different modes of operation. Two network management functions are discussed: call admission ensures a preset quality-of-service requirement by blocking new calls that causes congestion; route-configuration assignment determines the multicast tree for distributing the video of the current speaker. The reroutable route-configuration assignment is introduced. It allows a change of route-configuration when there is a change of speaker. Two reroutable assignment schemes are studied. In the normal scheme, a conference is always rerouted to the least congested route-configuration; while in the sticky scheme, a conference is only rerouted when the current route-configuration is congested. The video freeze probability, rerouting probability and the extended capacity space are derived. An example shows that the video freeze probabilities of the two schemes do not differ significantly. The sticky scheme, however, is superior as it gives a much smaller rerouting probability than the normal scheme. © 1998 John Wiley & Sons, Inc.

A (p,k) ShuffleNet is a type of regular multihop network with kp/sup k/ nodes. If only some of the nodes are extraordinarily busy, these so-called active nodes can be assigned to specific ShuffleNet locations to minimize the average hop count. An exhaustive search for the optimal node placement is not feasible for any reasonable size networks, particularly for networks requiring frequent reconfigurations, i.e. adding and dropping active nodes and changing traffic rates. The authors propose a computationally efficient algorithm that can give near optimal solution to the above problem. The procedures of adding and dropping of active nodes are also described.>