In order to use WDM networks efficiently, a lightpath is dynamically established upon the arrival of a connection request. Dynamic routing is usually determined by conventional shortest path algorithm. Different link weight assignment results in different routing and network performance. In this work, we propose a new weight assignment scheme in which link weights are proportional to the exponential of link utilization. It is used to differentiate network in different traffic loads so as to arrive load balanced routing. Based on the assignments, one adaptive routing algorithm referred to as Exponential Weighted Shortest Feasible Path (EWSFP) is presented. Furthermore, to avoid consuming too much network resources by a path with large hop count, a new call control algorithm named Random Early Blocking (REB) is proposed. By randomly rejecting calls with long routing path in heavy load, the whole network blocking probability can be greatly improved. Moreover, to balance the favor of short distance calls, a fair version of REB (called FREB) is presented in the paper. Extensive simulations are performed to evaluate the performance among different schemes. Compared to well-known shortest feasible path routing and weighted feasible shortest path routing, EWSFP achieves lowest call blocking probability. In cooperation with REB, all of the three adaptive routing approaches can significantly improve their performance. Among them the EWSCP+REB scheme outperforms all of the others. Simulation results also reveal that the FREB can obtain fairness among calls with different distances by only sacrificing a little blocking performance.
