The continual growth of Internet traffic necessitates a high-capacity transmission platform and also highlights the importance of a multi-granularity transport network due to the flexible bitrates of Internet traffic. To meet these challenges, a spectrum-flexible optical network based on orthogonal frequency-division multiplexing (OFDM) transmission technologies has been proposed as a promising solution because of its large transport capacity and bandwidth flexibility. In spectrum-flexible optical networks, we focus on the complete dynamic routing and spectrum assignment (RSA) problem, which integrates the signal format selection subproblem, the routing subproblem, and the spectrum assignment subproblem. In the complete RSA problem, we jointly consider the spectrum continuity constraints, the transmission distance constraints, and the relationship between the traffic bitrate and the signal bandwidth. A nonlinear programing model is presented to state the complete RSA problem. To solve the problem, we introduce a decomposition approach that divides the nonlinear problem into three steps: selecting the modulation format, solving the linear basic RSA problem, and checking the transmission distances. Based on the decomposition, we propose two heuristic approaches: modulation level fixed and adaptive RSA approaches for the complete RSA problem. Through analysis of the approaches, we prove that, in theory, the modulation level adaptive RSA approach can find the optimal solution to the complete RSA problem. The two approaches are implemented and compared in simulations. The results prove that the modulation level adaptive approaches achieve a lower capacity blocking probability than the modulation level fixed approaches.
© 2012 OSAFull Article | PDF Article