跨越40 Gb/s之下一世代OFDMA/WDM混合型光纖接取網路：設計與實作

Abstract

目前Gigabit被動光網路(GPON)採用開關鍵控(OOK)調變格式來提供2.5 Gb/s下載速率與1.25 Gb/s上傳速率，以及採用分時多重存取(TDMA)提供最多64位終端使用者之頻寬分配。然而，隨著網際網路電視(IPTV)與高畫質影片(HD)之高速寬頻服務成長，未來頻寬在成本效益考量下將必要且適時地進展至10 Gb/s、40 Gb/s、甚至更高頻寬。展望未來，拜數位信號處理(DSP)技術發展之所賜，正交分頻多重存取(OFDMA)之PON網路已成為下一世代具有成本效益與靈活變通性質之被動光網路卓越候選者。隨著彈性使用較高階數正交振幅調變(QAM)技術，OFDMA-PON可有效地達到高光譜效率與盡量減少光元件頻寬需求。然而，OFDMA PON有諸多問題尚待解決，如光拍頻干擾(OBI)與高峰均功率比(PAPR)等，嚴重減損其傳輸表現。這些問題需要新的PON系統架構與傳輸技術設計來解決。 另一方面，分波多工(WDM) PON亦被廣泛期待為諸多具前景性之未來光接取系統之一。WDM PON可提供給每個ONU單一波長通道，其在遠端採用WDM多工器與解多工器來取代被動式分光器。然而WDM PON同樣面臨幾項嚴重的經濟與技術障礙。首先為了達到ONU可以不受特定光波長之限制使用，ONU需要使用可調式傳送器或是無色光源技術。然而寬頻可調式雷射目前極其昂貴。無色光源元件如反射式半導體光放大器(RSOA)與Fabry-Perot依舊不夠成熟造成嚴重的頻寬縮減。更且，此無色型WDM PON技術無法達到次波長粒度之動態頻寬分配，使得頻寬分配彈性不足。 在這三年期計畫中，我們在後卓越計畫已充分合作之三位教授再次組成一個研究團隊，最終目標在實現設計、分析與原型創建一個具高度擴展性之跨越40 Gb/s混合型OFDMA/WDM PON。這個混合型網路將擷取OFDM與WDM PON雙方系統之優點，且同時從架構與技術層面減緩兩邊之弱點。如下面時程表所示，這項計畫之顯著研究主題包含新OFDMA-based PON與混合型OFDMA/WDM PON架構設計、混合型媒介存取控制(MAC)與高彈性頻寬分配機制之設計與分析，以及每個光波長40 Gbps傳輸之原型實驗系統。

Currently, gigabit passive optical network (GPON) offers downstream (upstream) bit rates of 2.5 Gbit/s (1.25 Gbit/s) based on the on-off keying (OOK) modulation format and time division multiple access (TDMA) allowing the bandwidth to be shared by no more than 64 end-users. However, with the rapid growth of high-speed services like Internet protocol television (IPTV) and high-definition (HD) video, further bandwidth increases cost-efficiently to 10 Gb/s, 40 Gb/s, and beyond become essential and timely. Thanks to the advance in digital signal processing (DSP) technology, orthogonal frequency-division multiple Access (OFDMA)-PON has been envisioned as a prominent candidate for the next-generation cost-effective and flexible PONs. With the flexibility of using higher order quadrature amplitude modulation (QAM), OFDMA-PON effectively achieves high spectral efficiency and ultimately lowers the bandwidth requirement of components. However, OFDMA PON has several problems, such as the optical beat interface (OBI) problem and high peak-to-average power ratio (PAPR), culminating in serious degradation of transmission performance. As a result, such problems bring about the necessity of designing new PON system architectures and transmission technologies. On the other hand, wavelength division multiplexing (WDM) PONs have also been widely expected to become one of the most promising future optical access systems. A WDM PON provides a single wavelength channel to each ONU by using a WDM multiplexer/demultiplexer (instead of the passive optical splitter) at the remote end. However, WDM PONs also face several serious economic and technological hurdles. First, to make ONUs free from the restricted use of specific wavelengths, ONUs at WDM PONs are required to either use either tunable transceivers or to be color-less. However, broadband tunable laser is currently very expensive. The color-less devices, such as reflective SOA, and Fabry-Perot, are far from being mature, resulting in severe reduction in bandwidth. Moreover, such color-less-based WDM PON technology fails to achieve dynamic bandwidth allocation with sub-wavelength granularity, resulting in significant reduction in bandwidth allocation flexibility. On this three-year project, we (three professors) team up again (after post-Excellency project) to achieve the ultimate goal of designing, analyzing, and prototyping a highly scalable hybrid OFDMA/WDM PON of 40 Gb/s-and-beyond. The hybrid OFDMA/WDM PON takes advantages of the strengths of both OFDM and WDM PON systems, and at the same time mitigates their weaknesses from both architectural and technological perspectives. As shown in the roadmap, prominent research topics in this project include architectural designs of new OFDMA-based PON and hybrid OFDMA/WDM PON, design and analyses of hybrid medium access control (MAC) and high-flexibility bandwidth allocation schemes, and significantly the prototyping of 40Gbps-per-wavelength system.