High-Speed PAM-4 Signal Transmission For Data-Center Photonic Interconnections

With rapid development of online social networks,real-time video and other end-user services,as well as cloud computing,cloud storage,cloud platform and other business services,there occurs an urgent requirement of transmission capacity enhancement for data-center photonic interconnections(DCI).Due to the DCI characteristics of high speed,low cost and low power consumption,direct detection transmission system supported by advanced modulation formats together with digital signal processing(DSP)becomes the capable candidate for DCI applications.However,current nonlinear equalizers,with extremely high computational complexity during the DSP implementation,is the main obstacle.Therefore,my thesis focus on the research from the perspective of the DCI transmission model of high-speed PAM-4 signal.Then,the transmission impairments compensation methods based on DSP is comprehensively investigated,and the solutions with high performance and low computational complexity are verified.The main research outcomes of the thesis include:(1)The transmission system structure and transmission model of DCI are proposed.In particular,five sub-models constitute the transmission model are investigated in details,including the optoelectronic devices bandwidth limitation model,the optoelectronic devices noise model,the chromatic dispersion(CD)model,the fiber nonlinearities model,and the signal-to-signal beating noise(SSBN)model.Then,the signal distortion induced by transmission impairments in each model is analyzed and visually simulated.Finally,the mathematical representation of each transmission impairment is investigated,and we identify that the signal distortion induced by optoelectronic devices limited bandwidth,SSBN and fiber nonlinearities are related to the linear function,quadratic function and cubic function of the signal.(2)In order to solve the problem that the computational complexity of the Volterra equalizer commonly used in current DCI is too high,we propose the sparse Volterra equalizer.Based on the sparsity of Volterra kernels,sparse Volterra equalizer is implemented by modified Gram-Schmidt orthogonal with re-orthogonalization algorithm and orthogonal search algorithm.Finally,experimental verification of O-band 2 × 64 Gb/s PAM-4 signal transmission over 70 km standard single mode fiber(SSMF)is successfully realized.The sparse Volterra equalizer can achieve similar system performance but with half the computational complexity in comparison with the traditional Volterra equalizer.(3)In order to further reduce the computational complexity of the Volterra equalizer from the physical essence,we propose the digital CD pre-management technique,based on the discovery that the self-phase modulation(SPM)induced impairment can be optically mitigated by the residual positive chromatic dispersion of the SSMF link.Finally,experimental verification of C-band single lane 112 Gb/s PAM-4 signal transmission over 80 km SSMF is successfully realized.We find that with the help of digital chromatic dispersion pre-management for the purpose of both CD compensation and nonlinear impairment mitigation,the simplified Volterra equalizer without 3rd-order kernels is good enough to compensate the various transmission impairments and reach the 7% hard-decision forward error correcting threshold successfully.