Study On Forward Error Correction And Other Key Techniques Of10G Ethernet Passive Optical Networks

With the rapidly-increasing bandwidth requirement of users and development of application diversity, optical access networks have moved into the development stage of10Gbps bandwidth. Due to its high bandwidth, low cost, and stable performance,10G EPON has become one of the main optical access techniques. However, the10G EPON still faces several technical problems which require to be solved immediately including economic and efficient improvement of system capacity, significant upgrade of service throughput, and accurate and effective identification of applications. Therefore,10G EPON has become one of the research focuses all over the world.In order to solve the above problems of10G EPON, this thesis mainly studies on the key techniques of10G EPON including FEC, DBA, deep identification of applications. A series of innovative research achievement have been obtained. The main research work and innovative contribution are as followings.(1) In currently-used10G EPON, FEC uses static code without consideration of channel differences of optical network units, thus introducing waste of parity octets. In order to solve this problem, a novel channel-adaptive FEC scheme using variable codes is proposed. Based on the simulation results, it is shown that compared with static-code FEC, the proposed channel-adaptive FEC can improve system capacity by7%and be compatible with the specification of the standard. The other reason that causes inadequate utilization of FEC parity in10Gbps PON is the application unawareness and equal protection. In order to solve this problem, an application-adaptive FEC that adjusts FEC ratios according to the differential importance of applications has been put forward. Based on simulation results, it is indicated that the proposed scheme can achieve12%improvement of system capacity; meanwhile, it can be implemented with low increment of hardware complexity and be compatible with the specification of the standard.(2) To solve the problem that upstream TCP throughput et al. of PON cannot achieve the optimum level, which is caused by independent operation of TCP in transport layer and DBA in RS layer, a double-phased polling DBA with optimum polling cycle (OPCDPP-DBA) is proposed. In this scheme, the polling cycle is adaptively adjusted based on the average TCP transmission window. It is shown by simulation results that compared with isolated-operating DBA, OPCDPP-DBA can improve upstream TCP throughput of PON by at least20%, and the proposed DBA scheme decreases the hardware complexity of DBA by50%.(3) In order to achieve an effective foundation and premise for service carrying strategy, a deep application identification scheme for10G EPON is proposed based on the concatenation of DPI and DFI, and bidirectional coordination of OLT and ONU. It is shown by the simulation results that the proposed scheme can gain at least90%identification accuracy ratio and recalling ratio and achieve real-time operation at lOGbps.(4) A10G EPON system scheme which supports adaptive FEC, cross-layer DBA, and deep application identification is proposed. Moreover, on the issue of adaptive FEC, a scheme of variable-code encoder and decoder is put forward based on reconfiguration and modification and implemented on FPGA. FPGA synthesis results show that this scheme can achieve at least25%decrement of hardware complexity for adaptive FEC at10Gbps.