By Naveena Genay, Benoît Charbonnier, Emmanuel Grard, and Pierre Wolkowicz
Orthogonal frequency-division multiplexing (OFDM), leveraging technology developed for 10-Gbps metro networks, offers an interesting option for high bit rates and reach extension in next-generation optical access networks.
The use of broadband connection technologies is growing strongly around the world. Nevertheless, the emergence of new services—ultrahigh-definition (HD) video, local and global storage-area networks, file sharing, HD video-on-demand, video conferencing, interactive online gaming, home networks for in-home digital cinema—threatens to create a bandwidth bottleneck. For this reason all of the major operators say that the age of FTTH has finally arrived.
Yet FTTH must continue to evolve to address future bandwidth demands. This evolution likely will see the implementation of new network architecture like TDM or WDM-PON.
On the other hand, intensive research work reported recently at major optical telecommunications conferences like OFC/NFOEC and ECOC has focused on new modulation formats with very high spectral efficiency, especially modulation based on orthogonal frequency-division multiplexing (OFDM). With these new modulation techniques, innovative equipment will be able to overcome bandwidth bottlenecks when high-bit-rate services become massively in demand.
FTTH technology alternatives
According to the forecast shown in Fig. 1, fiber-optic-based broadband services will account for 23% of the total broadband connections in the 10 countries studied. While mobile broadband is growing substantially (3G, WiMAX, etc.), fixed broadband will keep its technology advantage in regard to higher available bandwidth.
The options commonly available today, such as GPON, are based on TDM. On the other hand, alternatives based on WDM are emerging. WDM-PON is expected to provide a dedicated 100-Mbps fiber connection to homes and businesses. In this architecture, Fabry-Perot laser diodes are used as sources, with injection locking by means of filtered amplified spontaneous emissions (ASE) from a broadband light source such as a pumped erbium-doped fiber.
Though the system is cost-effective, the bit rate achieved is about 1 Gbps over a distance of 20 km. The only deployments of WDM-PON in the world are in Korea. This approach also has not yet been standardized by the ITU.
CWDM-PON also offers a low-cost alternative, as opposed to DWDM where the spacing between each wavelength is 20 nm. Other emerging technologies are based on code-division multiplexing (CDM). However, for the moment the bit rate per user is limited to 100 Mbps.
Nevertheless, a migration of TDM-PON toward hybrid WDM/TDM-PON seems more realistic since it allows more bandwidth sharing.
The IEEE and Full Service Access Network (FSAN) groups are actively involved in developing standards for next-generation access networks and are working toward the standardization of 10-Gbps systems. The IEEE released a first draft of the 10G EPON specifications in November 2007 via the IEEE 802.3av task force.
The cost of transmitters and receivers is an issue in access networks since the overall network cost is driven by the number of users. Transmitters and receivers for 10 Gbps exist for long-haul network applications; however, their cost due to very challenging performance requirements is not low enough for the access market.
On the other hand, an OFDM technique—which is derived directly from the radio transmission domain and implemented in xDSL transmission—offers a potentially low-cost option for two main reasons. First, the technology derived from the development of linear distributed feedback (DFB) lasers used in radio transmission can be directly applied to the access network and this can help to reduce the cost. Laser and receiver module technologies developed to support the 10-Gbps digital metropolitan market have now reached enough industrial and cost maturity to be seen as interesting technologies that can be adapted to support the access market. Secondly, the OFDM technique is mature, and so the development of associated software will not be a barrier.
Recent progress in the field of optical sources and filtering components enables the consideration of approaches involving source sharing over a PON based on WDM. Moreover, with progress in digital communications, an increase in the bit rate in the access network may be possible.
OFDM and its advantages
OFDM is a multicarrier technique where a high-speed data stream is divided into multiple parallel lower-speed streams and modulated onto subcarriers of different frequencies for transmission (see Fig. 2). Usually, each subcarrier is mapped onto complex values according to m-ary quadrature amplitude modulation (m-QAM) constellation mapping. In practice, the fast Fourier transform (FFT) algorithm is used to generate and demodulate OFDM signals.
OFDM improves the robustness of high-bit-rate signals against adverse effects such as chromatic dispersion, which can cause intersymbol interference. Indeed, splitting the signal into lower-bit-rate signals per subcarrier and using a cyclix prefix counteracts the penalties induced by the optical fiber.
OFDM is a good candidate for implementation in new systems to meet the requirements of future optical access networks in terms of extended reach and bit rate. This format is appropriate to meet the constraints brought about by the convergence of metro and access targeted by long-reach PONs.
Indeed, the well-known classical NRZ format has shortcomings when higher bit rates such as 10 Gbps or more are required. In-line dispersion compensation could be cost-prohibitive, and complex engineering rules could arise and further raise the cost of optical access deployment. Another advantage of OFDM is the application of OFDMA to dynamically allocate subcarriers or services to users.
Different network architectures are targeted for OFDM implementation (see Fig. 3). These include TDM-PON and WDM-PON with a remodulation architecture that implements colorless optical networking units (ONUs). The innovation in these architectures lies in applying the OFDM modulation format by direct modulation of low-bandwidth lasers and direct detection through photodiodes. Also, OFDM format does not require dispersion compensation, which is another improvement versus well-known alternatives. Moreover, the coexistence of different modulation formats such as NRZ and OFDM in the same architecture is feasible.
Activity in this field
Work has begun all around the world to investigate the use of the OFDM format modulation to develop new optical network architectures.
For example, projects such as the French ANR project “Enhanced PON using OFDM Modulation Format” (EPOD) and the European project “Architectures for fLexible Photonic Home and Access networks” (ALPHA) have been launched recently to develop new systems to meet the requirements of future optical access networks in terms of extended reach and bit rate using OFDM (see Fig. 4). NEC Labs and the National ICT Australia have reported through several publications successful transmission of 10-Gbps OFDM signals.
To achieve the targeted system performance and especially to meet OFDM 40-Gbps requirements, some new analog emitters and receivers are being developed. These modules will have high linearity, high output power (>10 dBm), and high bandwidth (>10 GHz).
Recently, as a first experiment, Orange Labs demonstrated a transmitted bit rate as high as 19 Gbps over a 25-km optical link for next-generation PON (NG-PON) by direct modulation of a DFB laser module having only 7-GHz bandwidth. Moreover, a bit rate of 10 Gbps was achieved over 100 km using the same technique without chromatic dispersion compensation.
This technique, which is applied in xDSL transmission, enables high bit rates by adapting the bit and power level per subcarrier as a function of the frequency band. This is achieved by measuring the channel state information for each subcarrier. In our case, the CSI was the signal-to-noise ratio (SNR) measured at the receiver side. Consequently, subcarriers with a high SNR will be assigned more bits to transmit than subcarriers with a lower SNR.
Such preliminary results offer the promise of future groundbreaking capacity by direct modulation of high-bandwidth analog lasers to transmit a signal with a bit rate higher than 10 Gbps in the access network. Moreover, by considering the application of OFDM in WDM-PON or TDM-PON architectures, it may therefore be possible to overlay next-generation equipment capabilities on top of a previously installed optical access network such as GPON by just changing the terminal equipment.
Industrial maturity and the cost barriers associated with using laser and receiver modules developed for the 10-Gbps metropolitan market are becoming less of an issue in the evolution of optical access networks. As these networks push the limits of bit rate and span transmission extension, 10-Gbps metro laser and receiver technologies have a great opportunity in the next few years to expand their lifetimes in support of access network requirements at least down to the optical line terminal.
Naveena Genay andBenoît Charbonnierare research engineers at Orange Lab (Lannion, France). Emmanuel Grard is development program leader andPierre Wolkowicz is product line manager at 3S PHOTONICS (www.3sphotonics.com).
Links to more information
LIGHTWAVE: OFDM Promises 100-Gbps Alternative
LIGHTWAVE ONLINE: Ofidium Secures Funding for 100-Gbps Telecommunications
FRANCE TELECOM: Orange Labs in France