Assessment of linear semiconductor amplifiers for upgrading WDM-PONs with wavelength reuse Juan José Martínez,1,* Noemi Merayo,2 Asier Villafranca,1 and Ignacio Garcés1 1
Laboratory of Optical Transmission and Broadband Technologies (TOYBA), Photonics Technology Group (GTF) of the i3A, Walqa Technology Park Cuarte (Huesca) 22197, Spain 2
Optical Communications Group of the Department of Signal Theory, E.T.S.I. Telecomunicación, University of Valladolid, Paseo de Belén 15,Valladolid 47011, Spain *Corresponding author:
[email protected] Received 14 July 2011; revised 1 December 2011; accepted 7 December 2011; posted 7 December 2011 (Doc. ID 151123); published 13 February 2012
In this work we have assessed the capacity of a linear semiconductor optical amplifier to compensate the fiber and component losses present in a wavelength division multiplexing passive optical network (WDM-PON) evolution from fiber-to-the-building (FTTB) to fiber-to-the-home access. The evaluation measurements confirm that the presence of a semiconductor optical amplifier placed at the entry of a group of optical network units that share the same wavelength channel can raise the loss budget that the link can tolerate in the fiber, compensating for the losses of a passive splitter up to a 1:16 division rate, allowing the upgrade of existing WDM-PON FTTB structures to make the fiber reach the final user’s home. © 2012 Optical Society of America OCIS codes: 060.4510, 250.5980.
1. Introduction
Access networks are currently experiencing a profound change that is redefining the entire network architecture in order to keep up with the increasing bandwidth demand of subscribers while maintaining reasonable costs for the operators. As bandwidth requirements in the network surpassed the capacity of copper cables, optical fiber has been extending from the operator end closer to subscribers, and this tendency is already leading to all-fiber networks with fiber-to-the-home (FTTH) being intensely deployed at present [1,2]. But the growth of the optical access network is happening toward both ends. At the operator end, it is being proposed that metro networks can be reduced or even may no longer be needed by instead adopting a long-reach passive optical network (PON) to take the central-office (or optical line terminal, OLT) of the access network closer to the 1559-128X/12/060692-05$15.00/0 © 2012 Optical Society of America 692
APPLIED OPTICS / Vol. 51, No. 6 / 20 February 2012
core network and thus enabling physical merging of central offices (COs) with the consequent cost reduction. This evolution is not straightforward, as an important redesign of the optical access network demands great investments that require a clear return and a long prospective life span. Figure 1 presents a wavelength division multiplexing PON (WDM-PON) distribution with home users that receive the signal via a final copper distribution in a fiber-to-the-building (FTTB) model architecture. WDM-PONs present some advantages, including a symmetrical bandwidth or its inherent point-to-point topology. However the higher cost of this architecture makes it difficult to implement a FTTH model in the already established networks. Present solutions to obtain higher power budgets for long reach or higher density PONs often imply the use of external amplification by distributed Raman amplification or the presence of erbium-doped fiber (EDF) segments acting as an EDF amplifier (EDFA) in the link path [3–5]. Both solutions are effective, but the network has to be specifically built with these characteristics
Fig. 1. WDM network with copper distribution to the final users.
in mind: for the Raman amplification not all the fibers are equally suitable, while the EDF sections have to be previously introduced in the structure; also both of them need adequate pumping sources that have to be placed in the infrastructure and may cause cross talk problems if the channel distribution is not considered. For these reasons, in already deployed networks these alternatives are not the ideal solution. Our proposal, depicted in Fig. 2, uses a linearized semiconductor optical amplifier (LSOA) in the distribution cabinet inside a building to enable transition from FTTB to FTTH. Also the gain of the device can be used (compensating losses) to extend the PON range in order to adopt a long-reach PON architecture that reduces the size of the metro distribution, taking the CO closer to the core network. The presence of the LSOA allows raising the power budget in an already installed WDM-PON without making any major changes in the actual infrastructure. Also the smaller size on an LSOA compared with an EDFA makes the device easier to install in order to achieve a budget gain, and its more consistent response against the burst signal in a PON network makes the budget extension more stable [6]. In this paper, we will study the possibilities of network upgrading achieved with an LSOA device in a WDMPON architecture based on colorless reflective optical network units (ONUs) and analyze the possibility to use it to extend the available power budget. 2. LSOA Characteristics
The chosen amplifier is an LSOA (model SOA-LOEC-1550 from CIP) with a low confinement factor and optimized for low noise and linear behavior. Although the device has low gain (16 dB max), its characteristics of saturation and noise make the LSOA well suited for in-line amplifying tasks; also, its broadband characteristics make the device suitable for working in the whole C band. When various signals are amplified using a single SOA device, if one of the signals shows much higher optical power than the others (and if the device is operating in saturation), it can deplete the carriers of
Fig. 3. Output optical power curves for the LSOA with two inputs against different Ch 1 injected powers, for different Pch2 optical powers. The upper graphic shows the output for Ch 1, and the lower graphic shows the output for Ch 2.
the device active zone, effectively reducing the gain of the other signals. This nonlinear phenomenon is known as cross-gain modulation (XGM). Considering the LSOA function in the network (simultaneous amplification of upstream and downstream signals in a channel), it is important to determine its XGM effects; thus, a characterization has been performed in the LSOA with the results presented in Fig. 3. Two different signals have been injected at the same time in the device: one variable for each measurement (Ch 1) in one direction and the other fixed (Ch 2) in the other direction. Output power curves of Ch 1 show little difference with the change of the Ch 2 input power: in the worst case (Pch2 6 dBm), the gain has a penalty of no more than 2.5 dB. Also it is important to remark that the gain of the device only deviates from the linear at the highest Ch 1 input powers. The same behavior can be seen for the output powers of Ch 2, where in the worst scenario (an input power of −20 dBm), the compression of the output power of channel 2 is only around 3 dB (for a Ch 1 input of 7 dBm). The results show that the device is capable of working with different input optical powers without excessive XGM compression—compression that only appears at the worst cases. In addition, even for the compressed cases, the LSOA gain shows good values above 10 dB. Considering the results, the proposed LSOA is capable to work as an in-line amplifier for a PON network without XGM problems, although a custom more linear LSOA design could achieve even better behavior, with less XGM effect. 3. Network Architecture
Fig. 2. Proposed network extension using an LSOA.
The analyzed network architecture (Fig. 4) is an access WDM-PON working on the optical C band. It has a reflective ONU that modulates the optical 20 February 2012 / Vol. 51, No. 6 / APPLIED OPTICS
693
Fig. 4. WDM-PON architecture with one unique wavelength for both links, different quasi-orthogonal modulation schemes, and the presence of an LSOA.
source provided remotely by the OLT, so no active lasing devices can be found at the ONUs. The sources are different directly modulated DFB lasers (DMLs) used for both downstream and upstream communications [7]. Each DML works at a different wavelength and gives service to a different set of ONUs where the signal is divided using a standard optical splitter and shared via a time-division multiplexing (TDM) strategy; therefore, each subgroup that uses the same wavelength can be viewed as an individual TDMPON, and the entire network as a low-cost hybrid WDM/TDM-PON [8,9]. To achieve an effective separation of the data with the same source, two different quasi-orthogonal modulations schemes have been chosen: for the downstream, a narrow-frequency-shift keying (FSK) modulation obtained via the direct modulation of the DFB bias current (thanks to the laser intrinsic adiabatic chirp [10,11]), while the upstream uses a standard intensity modulation (IM) introduced via a modulated reflective semiconductor optical amplifier (RSOA). Both communications (up and down) work with a 1 Gb∕s Ethernet signal. In this environment, the LSOA is introduced before the splitter in front of each group of ONUs, typically in an indoor cabinet of a building. The topology depicted in Fig. 5 is an example of the device placed in a more real network [12], where the LSOA is used to compensate the losses of the last splitting (inside the building) where the electrical power is available and the splitting
Fig. 5. Placement of the LSOA in a realistic topology used to compensate the last splitting losses in the building. 694
APPLIED OPTICS / Vol. 51, No. 6 / 20 February 2012
ratio is not excessively elevated (1:4 or 1:8); given the reduced size of the device, it presents no problem for its installation in this environment [6]. 4. Experimental Measurements A. Experimental Setup
In previous works we demonstrated that the effects of the fiber dispersion over a 1 Gb∕s Ethernet transmission are hardly noticeable, even over distances of 50 km [7]. The possible backscattering penalties that could arise in the fiber as a consequence of using a single wavelength for both upstream and downstream channels have been established as negligible in previous experiments with a similar architecture [13], as long as the injected optical power is maintained in low enough levels (below 7 dBm). So, excluding these other possible penalty effects, the main focus of the experiments is to know how the total link loss can affect the transmission. The experimental setup used to determine the performance of a single link is shown in Fig. 6. To simulate an OLT of the architecture in Fig. 4, we use a DFB laser centered at 1550.1 nm (output power of nearly 4 dBm), its bias current is directly modulated with a small excursion (Ibias 75 mA 8 mA) in order to obtain the desired FSK modulation; the circulator is used to discriminate incoming from outgoing signals at the OLT. The optical variable attenuator simulates the fiber and other component losses, while the filter represents the WDM demultiplexer that delivers the individual wavelengths to each group of users; finally, after the filter we place the LSOA. Between the amplifier and the ONU, a fixed (13.5 dB) optical attenuator is situated in order to represent a nonideal 1:16 power splitter (12 dB if ideal). At the ONU, the received signal is divided and filtered: one part is redirected to the RSOA, modulated, amplified, and sent back to the OLT,
Fig. 6. (Color online) Experimental setup to evaluate the performance of the link against the attenuation budget.
conforming the upstream link; while the other part passes through a demodulator filter that suppress one of the downstream FSK peaks, obtaining an amplitude modulation that can be directly detected for the downstream signal. The modulations source is a gigabit Ethernet (GbE) optical traffic analyzer. B.
Results
Measurements have been made with and without the combination of the LSOA and attenuator (in the dotted square of Fig. 6). Link quality is evaluated considering the bit error rate (BER) of the received signal against the total losses in the link path: only the variable attenuator in the case without LSOA, the variable plus the fixed attenuator in the cases with amplification. The curves in Fig. 7 show the results: downstream performs better than upstream for both amplified and nonamplified cases, which makes the upstream link the most limiting case. This higher penalty in the upstream channel is caused by two elements: the residual amplitude modulation from the downstream narrow-FSK that affects the IM upstream reception (upstream and downstream communications are not perfectly orthogonal) and noise caused by the amplifiers that have a more significant negative
effect over the IM link [7]. In the measurements without the LSOA, the downstream link can tolerate (for an error-free transmission) a link loss of 22.5 dB; for the upstream link, this margin is reduced to 16 dB, which is enough to overcome a mid-distance link and its additional components but cannot cope with a 1:16 (12 dB) splitting ratio to achieve an FTTH solution. On the other hand, the cases with the LSOA using the same receptors can tolerate a higher total link loss (in this case, the variable optical attenuator plus the fixed one)—up to 34.5 dB for the downstream and around 29.5 dB for the upstream. This allows compensating the 1:16 splitter (represented by the fixed attenuator). The amplifier device can provide enough margin for a budget compensation, making possible the upgrade of an existing network. One of the particulars in PON architectures is the burst traffic present in the upstream direction, causing differences in upstream optical power that enters in the LSOA and leading to link penalties. Different measurements have been made under burst conditions in order to evaluate this possible error source. Results show identical link performance for burst and no-burst upstream due to the linear characteristics of the LSOA: the gain of the amplifier does not
Fig. 7. Link performance for downstream and upstream links, with and without the use of the LSOA. 20 February 2012 / Vol. 51, No. 6 / APPLIED OPTICS
695
suffer variations caused by the different optical powers of a burst signal.
2.
5. Conclusions
It has been shown that the presence of an LSOA device can raise the available power budget in an already deployed network, which has been used to upgrade a WDM-PON access network to FTTH architecture without important alterations in its infrastructure. Also, the extra budget can be used to move the CO closer to the core network. The presence of the LSOA can help to upgrade a pure TDM-PON to a hybrid WDM/TDM-PON, compensating with the LSOA the increasing loss from the introduction of the arrayed waveguide grating and the in-building distribution. Also it can be used to compensate losses in a problematic network: rising the budget with a small device (that can be carried by the technician and installed on the spot). The solution also has flexibility advantages, because each individual wavelength can be upgraded independently. Another advantage for the practical implementation of the proposal is that LSOA devices could be easily integrated (resulting in a smaller, more efficient device), which can greatly reduce the fabrication cost, especially if the LSOAs are manufactured in elevated quantities (for an operator). Integration also permits having more control over the characteristics of the device (compared with the commercially available LSOA used in this paper) making it possible to obtain higher gain and better linear characteristics. This will allow obtaining a better amplifier that also can be specifically tailored to cover more precise requirements. This work was funded in part by the ARAID Foundation and Ibercaja within the program “Empresa Innovadora 2010” and by the Spanish MICINN by means of grant TEC2010-19418. References 1. F. Saliou, P. Chanclou, F. Laurent, N. Genay, J. A. Lazaro, F. Bonada, and J. Prat, “Reach extension strategies for passive
696
APPLIED OPTICS / Vol. 51, No. 6 / 20 February 2012
3.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13.
optical networks [Invited],” J. Opt. Commun. Netw. 1, C51–C60 (2009). M. D. Vaughn, D. Kozischek, D. Meis, A. Boskovic, and R. E. Wagner, “Value of reach-and-split ratio increase in FTTH access networks,” J. Lightwave Technol. 22, 2617–2622 (2004). I. Tafurmonroy, R. Kjar, F. Ohman, K. Yvind, and P. Jeppesen, “Distributed fiber Raman amplification in long reach PON bidirectional access links,” Opt. Fiber Technol. 14, 41–44 (2008). J.-P. Blondel, F. Misk, and P. M. Gabla, “Theoretical evaluation and record experimental demonstration of budget improvement with remotely pumped erbium-doped fiber amplification,” IEEE Photon. Technol. Lett. 5, 1430–1433 (2002). J. Lázaro, J. Prat, P. Chanclou, T. Beleffi, A. Teixeira, I. Tomkos, R. Soila, and V. Koratzinos, “Scalable extended reach PON,” in Conference on Optical Fiber Communication/ National Fiber Optic Engineers Conference, 2008 (IEEE, 2008), pp. 1–3. D. R. Zimmerman and L. H. Spiekman, “Amplifiers for the masses: EDFA, EDWA, and SOA amplets for metro and access applications,” J. Lightwave Technol. 22, 63–70 (2004). J. J. Martínez, J. I. Garcés Gregorio, A. López Lucia, A. Villafranca Velasco, J. C. Aguado, and M. Á. Losada Binué, “Novel WDM-PON architecture based on a spectrally efficient IM-FSK scheme using DMLs and RSOAs,” J. Lightwave Technol. 26, 350–356 (2008). F.-T. An, K. S. Kim, D. Gutierrez, S. Yam, E. Hu, K. Shrikhande, and L. G. Kazovsky, “SUCCESS: a nextgeneration hybrid WDM/TDM optical access network architecture,” J. Lightwave Technol. 22, 2557–2569 (2004). C. Bock, J. Prat, and S. D. Walker, “Hybrid WDM/TDM PON using the AWG FSR and featuring centralized light generation and dynamic bandwidth allocation,” J. Lightwave Technol. 23, 3981–3988 (2005). S. Kobayashi, Y. Yamamoto, M. Ito, and T. Kimura, “Direct frequency modulation in AlGaAs semiconductor lasers,” IEEE J. Quantum Electron. 18, 582–595 (1982). A. López, I. Garcés, M. A. Losada, J. J. Martinez, A. Villafranca, and J. A. Lázaro, “Narrow-FSK optical packet labeling scheme for optical ethernet networks,” IEEE Photon. Technol. Lett. 18, 1696–1698 (2006). T. Rokkas, D. Katsianis, and D. Varoutas, “Techno-economic evaluation of FTTC/VDSL and FTTH roll-out scenarios: discounted cash flows and real option valuation,” J. Opt. Commun. Netw. 2, 1042–1051 (2010) J. J. Martínez, I. Garcés, A. López, A. Villafranca, and M. A. Losada, “Analysis of the influence of backscattered optical power over bidirectional PON links,” Opt. Commun. 283, 2243–2250 (2010).
