Planning of transport networks based on photonic and electronic cross-connection

This paper deals with the problem of dynamic allocation of bandwidth in a h ierarchical transport platform comprised of both broadband (optical) and wi deband (ATM) crossconnects. The crossconnect platforms will react to fluctu ations of traffic on different time scales. The model developed in this pap er will take advantage of both multi-service and multi-hour character of tr affic profiles and includes off-line and on-line algorithms. The off-line m odel will generate the configuration of optical bearers and the routing of origin-destination paths in the network for each time interval. The on-line algorithms will adjust the capacity allocated to each path in accordance w ith the short-term fluctuations of demand to decrease blocking probability and maximize resource utilization. We formulate the problem of designing a minimum cost network for a multi-ho ur, multi-service traffic profile where the cost is a piecewise linear func tion of the maximum possible number of optical paths as well as a linear fu nction of the number of ports and of the traffic flow through tandem nodes. The model is a large-scale mixed integer program. Further, we provide a se cond model to carry out the fine tuning necessary to cater for traffic fluc tuations. The alternative versions of the second model are given, including a goal program that optimally matches the network capacity to the varying traffic requirements.