Transponder architecture supporting multi-wavelength source

Multi-flow transponder in elastic optical networks has been recently proposed as a flexible class of transponder that is able to generate multiple optical flows. Multi-flow transponders may support slice-ability, i.e. flows can be flexibly associated to the incoming traffic requests and they can be directed toward different destinations. Alternatively, flows can be co-routed. Transponders supporting slice-ability are also called sliceable bandwidth variable transponders (SBVT). Typically, in the literature, SBVTs have been considered composed of multiple laser sources (i.e., one for each optical flow unit, or sub-carrier). In this paper, we propose a novel cost-effective SBVT architecture based on multi-wavelength source and micro ring resonators (MRRs), supporting super-channel (i.e., an optical connection composed of several co-routed sub-carriers) and slice-ability. A multi-wavelength source is a source able to create several optical sub-carriers from a single laser. Thus, to generate N sub-carriers, N − 1 laser sources can be saved. Then, MRRs are used to select each sub-carrier and direct it to the proper modulator. In this paper, we design and analyze MRR-based filters for sub-carrier selection through simulations. The proposed sliceable transponder also enables multi-rate, multi-modulation format with optical reach adaptation, and time frequency packing (an advanced transmission technique that is an evolution of the faster-than-nyquist). A specific design of a SBVT enabling an information rate of 400 Gb/s is presented considering standard 100 GbE interfaces. The adoption of multi-wavelength is analyzed in terms of constraints in the routing and spectrum assignment (RSA), especially in the case of slice-ability. In particular, multi-wavelength source may create sub-carriers that are contiguous in frequency, thus with possible limitations in the RSA. In this paper, the impact of such constraint is studied through network simulations. Results show that the proposed SBVT architecture is able to achieve low blocking and reducing the number of lasers installed in the network.