Space division multiplexed (SDM) elastic optical network (EON) is considered to be a promising scheme for large-capacity optical communication networks. The static routing, modulation, spectrum, and space allocation (RMSSA) in SDM-EONs with bundles of single-mode fiber is studied. Considering the computational complexity of resource allocation formulation, a path-based integer linear programming (ILP) formulation with fewer variables and constraints is modeled to solve the static RMSSA problem. Then a heuristic algorithm named local optimal RMSSA (LO-RMSSA) is proposed to be applicable in large-scale network scenarios. The calculated metrics are the maximum index of utilized frequency slots, the local spectrum resource utilization, and the runtime of algorithms. The results show that the proposed ILP model and LO-RMSSA algorithm get higher computational efficiency with other metrics no worse than the existing one.

A hybrid shaping scheme, carrier-less amplitude and phase modulation geometric-probabilistic hybrid shaping 16QAM (CAP-GPS-16QAM), is proposed for data center networks. The probabilistic shaping is achieved by applying the proposed probabilistic subset shaping to the constellation after geometric shaping (GS). An intensity modulation/direct detection experimental system of 5 GBaud is built to verify the proposal. The results show that the gains of the proposed scheme are 2.4, 2.6, and 3 dB compared with CAP-GS-16QAM, CAP-PAS-16QAM, and CAP-square-16QAM, respectively, at BER  =  10^−  3. Meanwhile, in terms of generalized mutual information, the proposal improves the performance by 0.078, 0.125, and 0.153 bits/symbol compared with CAP-GS-16QAM, CAP-PAS-16QAM, and CAP-square-16QAM, respectively.

Fiber optical parametric amplifiers (FOPAs) are regarded as one of the candidate optical amplifiers for future ultralong-distance, large-capacity, and high-speed optical fiber communication systems. FOPAs provide high signal gain and low noise figures compared with other optical amplifiers. However, they suffer from gain fluctuation, which restricts commercial application. Optimization of the dispersion parameters of highly nonlinear fibers (HNLFs) employed as gain media is essential because they have an immediate impact on the flatness of the amplifier. An optimization method for multiple HNLF segments using the differential evolution algorithm is proposed to minimize spectral gain variation. We theoretically yield an FOPA with an average signal gain of 20 dB and gain fluctuation <0.5  dB within 400 nm.

For future elastic optical networks, the narrow filtering effect induced by cascaded reconfigurable optical add–drop multiplexers (ROADMs) is one of the major impairments. It is essential to accurately estimate the filtering penalty to minimize network margins and optimize resource utilization. We present a method for estimating filtering penalty using machine learning (ML). First, we investigate the impact of ROADM location distribution and bandwidth allocation on the narrow filtering effect. Afterward, an ML-aided approach is proposed to estimate the filtering penalty under various link conditions. Extensive simulations with 9600 links are implemented to demonstrate the superior performance of the proposed scheme.