Traffic Grooming Techniques in Optical Networks

With the increase of the number of wavelengths per fiber, waveband switching has been proposed to decrease the number of switching ports in optical nodes. Another concept, that of Light trails, allows the intermediate nodes along a lightpath to access the wavelength channel, aiming at the reduction of the number of wavelengths. Both techniques apply traffic grooming on different levels of the WDM network. In this paper, we combine them and compare the two switching techniques: waveband switching lightpath (WBS-LP) and waveband switching lighttrail (WBS-LT). Auxiliary graph models (LP_AG/LT_AG) are proposed for WBS-LP/WBS-LT respectively. These two auxiliary graph models can exploit not only the wavelength resource in the fiber links but also the limited waveband ports resource inside the multi-granular optical cross connects (MG-OXC) nodes. The influence of network parameters, e.g. number of wavebands, ports and transceivers, is studied. The proposed algorithms are compared with shortest path (SP), least-weighted path (LW), and K-least-weighted path (KP) algorithms; numerical simulations show their better performance. For different algorithms, WBS-LT can have better blocking performance than WBS-LP especially when add/drop waveband ports are the critical resources.