Transmission of coexisting Orthogonal Frequency Division Multiplexing (OFDM)-baseband (BB) and multi-band OFDM-ultra-wideband (UWB) signals along long-reach passive optical networks using directly modulated lasers (DML) is experimentally demonstrated.
When optimized modulation indexes are used, bit error ratios not exceeding 5 × 10−4 can be achieved by all (OFDM-BB and three OFDM-UWB sub-bands) signals for a reach of 100km of standard single-mode fiber (SSMF) and optical signal-to-noise ratios not lower than 25dB@0.1nm. It is experimentally shown that, for the SSMF reach of 100km, the optimized performance of coexisting OFDM-BB and OFDM-UWB signals is mainly imposed by the combination of two effects: the SSMF dispersion-induced nonlinear distortion of the OFDM-UWB signals caused by the OFDM-BB and OFDM-UWB signals, and the further degradation of the OFDM-UWB signals with higher frequency, due to the reduced DML bandwidth.
© 2011 Optical Society of America
In recent years, the development of optical access networks allowed introducing a multitude of new services for end customers that were seen as too futuristic just a decade ago. Most operators have been very successful in offering triple-play services to customers and look for equal success in offering quad and quintuple-play (5-play) services in the near future [1, chaps. 2–4], . These services include, among others, standardized wireless signals, like ultra wideband (UWB), for higher capacity wireless services, and Gigabit-Ethernet (GbE) service [1, chaps. 2–4], .
Long-reach (LR) passive optical networks (PONs) have been a hot research topic . They intend to extend the coverage span of PONs from the traditional 20km to around 100km by using optical amplification [3, 4, 5] and, eventually, wavelength division multiplexing (WDM) technology [4, 5, 6]. WDM-PONs allocate a different wavelength to each optical network unit (ONU), providing a separate point-to-point connection between each ONU, at users premises, and the optical line termination (OLT), at the central office (CO). In order to support 5-play services at lower cost, LR-PONs need to be transparent to wireless services and standards, and accept wireless already-standardized equipment [1, chaps. 2–4], . In order to comply with such requirements in a cost-effective way, orthogonal frequency division multiplexing (OFDM) has been envisioned as a promising modulation format for such LR-PONs because new standardized broadband wireless technologies — worldwide interoperability for microwave access (WiMax), UWB and long term evolution (LTE) — use OFDM signals and its variants. This fact enables the required optical-wireless convergence in an easy manner by transmitting the wireless services in their standardized frequencies and formats along the LR-PON without any transmodulation, allowing simpler and lower cost ONUs that only photodetect, filter, amplify, and radiate directly to establish the wireless connection, therefore avoiding the need of frequency up-conversion at user’s premises . In addition, the usage of the OFDM format for provision of the GbE service allows taking advantage of high spectral efficiency, which may be used to increase the data capacity if required.
This use of OFDM-based services in LR-PONs, allowing each service being delivered in its own standardized band and OFDM-based signal format, is the service provision approach followed in this work. Unlike, other authors have proposed the use of OFDM signals in LR-PONs to deliver a single data stream that time-multiplexes all provided services [5, 8]. A scheme more similar to our approach, in the sense that OFDM-based wireless and wired services are provided in different bands, has been proposed in ; however, a 40GHz millimeter wave is used to provide the wireless service which does not fall in any of the above-mentioned features of our approach.
In fact, the approach followed in this paper allows deploying wireless signals without additional cost compared to a conventional xPON network with an wireless transmitter/receiver located at the ONU side (as only one transmitter/receiver pair is required in both cases). As a result, operators have also to evaluate the requirements for extending protocols commonly used for wireless personal area networks (e.g. UWB, among others) considering transparent transmission along LR-PONs versus the advantages of managing all communication channels in the CO. As an example, for the successful deployment of such protocols, changes in the standardized timing parameters of the physical layer have to be considered, as the main characteristic of personal area networks is the short transmission distance in opposition to transparent transmission along LR-PONs. Nevertheless, such discussion is beyond the scope of this paper.
In addition to the advantages of using OFDM signals, the use of low-cost optical transmitters in LR-PONs is crucial, because a major investment is the cost associated with the installation, operation and maintenance of optical transmitters at the users’ premises [10, 11, 12]. In order to reduce transmitter complexity and implementation cost , [13, p. 53], directly modulated lasers (DMLs) have been proposed to be used in LR-PONs, recently [11, 12]. Our previous work has focused on the use of DMLs to transmit multi-band OFDM-UWB signals along distances of the order of 100km , without investigating the degradation caused by the coexistence of other OFDM signals, namely a base-band (BB) signal for the transmission of non-wireless related information like the GbE service. This is a relevant issue in order to assess the feasibility of LR-PONs with 5-play capacity, using DMLs as optical transmitters.
In this work, to the best of our knowledge, the first experimental investigation on the optical transmission of coexisting OFDM-BB and multi-band OFDM-UWB signals along distances typical of LR-PONs, using DMLs, is presented. Considering optimized modulation indexes, a reach of 100km of standard single-mode fiber (SSMF) without using optical dispersion compensation is achieved with a bit error ratio (BER) below 10−3 for the worst-performing signal and an optical signal-to-noise ratio (OSNR) of 25dB measured in a bandwidth of 0.1nm (25dB@0.1nm).
2. Experimental setup
Figure 1 shows the experimental set-up used to demonstrate and analyze the main impairments of the optical transmission of coexisting OFDM-(BB+UWB) signals along LR-PONs. The experiment emulates the downstream transmission of one wavelength, from one OLT to one ONU, in the context of a WDM LR-PON. Although two solutions have been devised for UWB signals transmission, namely OFDM and impulse radio , the OFDM solution is considered in our study due to the features mentioned in section 1.
In the OFDM-(BB+UWB) modulator, the OFDM-BB and the composite OFDM-UWB signals are individually generated and combined. Quadrature phase-shift keying (QPSK) is used as modulation format of the information sub-carriers for both OFDM-UWB and OFDM-BB signals. As explained below, in the experiment, the composite OFDM-UWB signal is composed of three OFDM-UWB sub-bands. Each UWB sub-band signal is based on ECMA-368 standard  using 128 sub-carriers, and its bandwidth is 528MHz. The bit-rate of each OFDM-UWB sub-band is 640Mbit/s. The BB signal uses 256 sub-carriers, from which 172 for data, 5 for pilots and 79 with zeros, and its bandwidth is about 660MHz. The OFDM-BB signal bit-rate is 1.26Gbit/s which is adequate to provide the GbE service. In the OFDM-(BB+UWB) modulator, and in order to reduce aliasing, each OFDM-UWB sub-band signal is filtered by a low-pass filter (LPF) with a bandwidth of 264MHz, before up-conversion, while the OFDM-BB signal is filtered by a LPF with a −3dB bandwidth of 680MHz. The off-line generated OFDM-(BB+UWB) signals (at A - Fig. 1) are converted to the electrical domain by a Tektronix interleaving Arbitrary-Waveform Generator (AWG) 7122B operating at 20Gsamples/s in continuous mode. The amplitudes of the OFDM-(BB+UWB) signals are controlled by an electrical driver and filtered by an electrical filter to remove high frequency components before being added to the DML bias current (Ib) and injected into the DML (at B - Fig. 1).
The DML is a commercially available multi-quantum well distributed feedback cooled laser from FITEL, with threshold current of Ith = 8.1mA. A chirp parameter of 2.6 was extracted from experimental results using the method presented in Ref. . Experimental measurements of the laser intensity modulation (IM) response were conducted for different bias currents, in order to gain insight on the main DML response limitations and on the DML bias current that leads to the best performance. Figure 2 shows the measured IM response of the DML, for several bias currents. Figure 2 shows that, due to the limited bandwidth of the DML, the UWB sub-band that follows the third UWB sub-band centered at 4.49GHz may be significantly degraded. As a result, only the UWB sub-bands centered at 3.43GHz (sb#1), 3.96GHz (sb#2) and 4.49GHz (sb#3) are considered. Moreover, Fig. 2 shows that the bias current should be carefully selected. In fact, bias currents higher than 30mA, although lead to higher output power, degrade further the laser IM response for UWB sb#3 due to parasitic effects, when compared with the case of bias current of 30mA. On the other hand, bias currents lower than 30mA, although lead to an improvement of the laser IM response for UWB sb#1-sb#3, degrade the system performance due to higher clipping and lower DML output power. Experimental tests of system performance were also conducted for different bias currents and driving powers. These tests confirmed that the system performance is degraded whether a bias current different from around 30mA is used. As a consequence, Ib = 30mA was selected which leads to an output power around 6dBm, centered at 1553.33nm.
The optical path is composed by spools of SSMF in order to consider the attenuation and dispersion of typical LR-PONs, an erbium doped fiber amplifier (EDFA), a noise loader and an optical filter (OF) with a -3dB bandwidth of 0.4nm centered at 1553.33nm to limit the noise bandwidth before the photodetector (PIN). The PIN has a -3dB bandwidth of 14GHz and an optimized power level at its input of −2dBm. The PIN has a DC block incorporated at its output in order to remove the DC component of the photodetected current. An optical spectrum analyzer is used to measure the OSNR at C (Fig. 1) with a resolution bandwidth of 0.1nm. As the SSMF has a dispersion parameter of 17 ps·nm−1·km−1 at 1553.33nm and SSMF lengths between 0 and 100 km are considered in the experiments, the dispersion accumulated along the SSMF ranges from 0 to 1700ps·nm−1 (for 100km of SSMF).
With the purpose of investigating the impact of the added noise by LR-PONs on the network performance, OSNR ranging from 20dB to 30dB are considered. Such range was selected in order to accommodate the expected OSNR of 28dB@0.1nm for a typical WDM LR-PON infrastructure, and composed by: i) 100km of SSMF with an attenuation coefficient of 0.3dB·km−1; ii) a WDM multiplexer and a demultiplexer with 5dB insertion loss each; iii) an EDFA with noise figure of 6dB and; iv) the previous transmitter and receiver power levels indicated in this section.
After the photodetector, the OFDM-(BB+UWB) signals are acquired in real time (at D -Fig. 1), at 20 Gsamples/s, using a digital storage oscilloscope (DSO) 81204A from Agilent. A synchronization marker is emitted by the AWG to the DSO, at the start of each transmitted OFDM-(BB+UWB) signals sequence, in order to synchronize the acquisition of data at the DSO. At the OFDM-(BB+UWB) demodulator, the OFDM-BB and OFDM-UWB signals are separated and demodulated off-line. In the demodulation process, the OFDM-UWB signals are down-converted and filtered by a LPF with a −3dB bandwidth of 264MHz to reduce the noise power. The OFDM-BB signal is filtered by a LPF with a −3dB bandwidth of 680MHz. The performance is assessed measuring the BER according to the method described in detail in  which allows measuring low BERs without resorting to long measurement times.
3. Experimental results
3.1. System optimization and performance assessment
In this section, the performance is assessed when transmitting the OFDM-(BB+UWB) signals along a SSMF length of 100km, typical of LR-PONs. In order to maximize the performance, the modulation indexes of the composite OFDM-UWB signal and OFDM-BB signal, mUWB and mBB, respectively, are optimized. The modulation indexes of the three UWB sub-band signals are set equal and given by . mUWB and mBB are defined as , where t is the time, T is the measurement time duration, and ilaser(t) is the signal driving the laser when the composite OFDM-UWB signal and the OFDM-BB signal are singly transmitted, respectively.
Figs. 3(a)–3(d) show the BER as a function of mUWB and mBB, after transmission along 100km of SSMF with an OSNR of 25dB@0.1nm for, respectively, the OFDM-BB and the OFDM-UWB sb#1–sb#3 signals. Figure 3(e) shows the maximum BER among OFDM-BB and OFDM-UWB signals, for each (mUWB, mBB) pair. Figure 3(e) shows that the optimum performance is achieved for mUWB = 13% and mBB = 6.6%, leading to BERs of about 5 × 10−4, 2 × 10−5, 8 × 10−5 and 4 × 10−4 for the OFDM-BB and the OFDM-UWB sb#1–sb#3 signals, respectively. Although the best BERs are much higher than 10−12, those BER levels show that, when the modulation indexes are optimized, acceptable performance can be achieved when common forward error correction (FEC) algorithms, whose BER threshold is 10−3, are used.
Figure 3(a) shows that, after transmission of coexisting OFDM-(BB+UWB) signals along 100km of SSMF, mUWB variations around 13% lead to a very reduced impact on the OFDM-BB signal performance. This means that the OFDM-UWB sub-bands induce only weak intermodulation components (nonlinear distortion) on the OFDM-BB signal. On the contrary, Figs. 3(b)–3(d) show that mBB variations around 6.6% lead to a strong impact on OFDM-UWB sb#1–sb#3 performance. This impact is attributed to the intermodulation components induced on OFDM-UWB signals by the OFDM-BB signal after SSMF transmission, which cannot be suppressed by electrical filtering at OFDM-UWB demodulator. Therefore, results shown in Fig. 3 indicate that the optimized OFDM-BB signal modulation index is imposed by the intermodulation induced by the OFDM-BB signal on the OFDM-UWB sub-bands performance.
Additionally, Fig. 3(a) shows a monotonic BER improvement of the OFDM-BB signal with the increase of mBB. This means that the OFDM-BB signal is much more impaired by noise than by signal distortion. On the other hand, Figs. 3(b)–3(d) show that, for mUWB lower than 13%, an increase of mUWB leads to BER improvement of OFDM-UWB sb#1–sb#3 signals as the limiting effect is noise. For mUWB higher than 13%, an increase of mUWB leads to BER degradation of OFDM-UWB sb#1–sb#3 as the limiting effect is distortion induced by the UWB sub-bands. These impairments of OFDM-UWB signals were already reported and discussed in  where a composite OFDM-UWB signal was singly (not in coexistence with any other kind of signal) transmitted. Thus, results shown in Fig. 3 indicate that the optimized OFDM-UWB signal modulation index is imposed by the distortion induced by each UWB sub-band on itself and on the other UWB sub-bands.
Figs. 3(a)–3(e) show that the optimum performance of coexisting OFDM-(BB+UWB) signals is imposed by the OFDM-BB and the OFDM-UWB sb#3 signals as the performance of OFDM-UWB sub-bands degrades with the increase of the sub-band central frequency. This increasing degradation with the sub-band central frequency is attributed to the larger IM response fluctuations within each sub-band for sub-bands with higher frequency (see Fig. 2), which are associated with the reduced DML bandwidth. In section 3.2, the effects that impair the optimum performance will be further experimentally analyzed and discussed.
In order to assess the importance of the conditions used for modulation indexes optimization, Fig. 4 shows the BER vs the OSNR for the coexisting OFDM-(BB+UWB) signals transmitted along 100km of SSMF. The optimized modulation indexes for 100km of SSMF and an OSNR of 25dB@0.1nm (mUWB = 13% and mBB = 6.6%) are used. Figure 4 shows a monotonic BER improvement for all OFDM signals with the increase of OSNR. For OSNR = 30dB@0.1nm, BERs of about 4 × 10−6, 2 × 10−6, 2 × 10−5 and 1 × 10−4 are achieved for the OFDM-BB and the OFDM-UWB sb#1–sb#3 signals, respectively. In addition, a significant improvement is observed for the OFDM-BB signal, when compared with the OFDM-UWB sb#1–sb#3. This is in accordance with the results of Fig. 3(a) where it can be seen that, around mUWB = 13% and mBB = 6.6%, the OFDM-BB signal is mainly impaired by noise and not by intermodulation distortion. Therefore, an increase of OSNR is more beneficial for the OFDM-BB signal than for the OFDM-UWB signals. Additionally, Fig. 4 shows that the BER improvement of OFDM-UWB sb#1–sb#3, resulting from the OSNR increase, decreases with the increase of the OFDM-UWB sub-band signal frequency. As Figs. 3(b)– 3(d) show, this is due to the increase of OFDM-UWB signal distortion for increasing frequencies, which results partly from the reduced IM response bandwidth of the DML.
3.2. Analysis of system impairments
In section 3.1, we have seen that the DML-based coexisting transmission system is mainly impaired by intermodulation distortion, optical noise and reduced IM response bandwidth of the DML. Other effects, such as laser noise [18, pp. 150–151], , electrical receiver noise [18, pp. 167–172], and power fading  have been indicated as impairments for this kind of systems. In this section, we conduct additional experiments in order to investigate further the several impairments of transmission of OFDM-(BB+UWB) signals along the system.
In order to identify the main origin of intermodulation distortion between coexisting OFDM-(BB+UWB) signals, the following fact is used. The difference of performance between transmission in coexistence and single signal transmission is due to intermodulation distortion between signals. As power levels propagating along the SSMF are sufficiently low to consider linear propagation along the SSMF, this intermodulation distortion can be due to the following causes: i) electrical to optical conversion in the DML/optical to electrical conversion in the PIN which, in case of existing, can be observed in back-to-back operation, and ii) SSMF dispersion-induced distortion, which occurs only after transmission along the dispersive SSMF.
Figure 5(a) shows the BER vs the SSMF length for the coexisting OFDM-(BB+UWB) signals and OSNR= 25dB@0.1nm. The optimized modulation indexes for 100km of SSMF and OSNR of 25dB@0.1nm (mBB = 6.6% and mUWB = 13%) are used. Additionally, Fig. 5(b) shows the BER vs the SSMF length when the OFDM-BB and the OFDM-UWB sb#1–sb#3 signals are singly transmitted with mBB = 6.6% and . The comparison of Fig. 5(a) with Fig. 5(b) shows that, in back-to-back operation, negligible differences are observed in the BERs of the signals, regardless they are singly transmitted (Fig. 5(b)) or transmitted in coexistence (Fig. 5(a)). This allows concluding that the intermodulation distortion between signals due to electrical to optical conversion in the DML/optical to electrical conversion in the PIN is negligible.
Figs. 5(a) and 5(b) show also that remarkable differences are observed between the BER of the different signals, in back-to-back operation. To explain these BER differences, Fig. 6 shows the electrical spectra at PIN output (at D - Fig. 1), in back-to-back operation, of the coexisting OFDM-(BB+UWB) signals for mBB = 6.6% and mUWB = 13% (Fig. 6(a)), and of the OFDM-BB and the OFDM-UWB sb#1–sb#3 signals singly transmitted with mBB = 6.6% and mUWB;sb = 7.5%, respectively (Fig. 6(b)–6(e)). The comparison of Fig. 6(a) with Figs. 6(b)–6(e) shows that similar spectra for OFDM-BB and OFDM-UWB signals are observed whether the signals are singly transmitted or in coexistence. This reinforces the conclusion that no significant intermodulation distortion between signals due to the electrical-to-optical and optical-to-electrical conversions is induced on the coexisting OFDM-(BB+UWB) signals. However, Fig. 6(a) shows that the OFDM-UWB sub-bands signals spectra are increasingly weakened and distorted for increasing frequencies. This is due to: i) the reduced IM response bandwidth of the DML, and ii) the fluctuations of IM response within the UWB sub-bands. This explains the highest BER of 6 × 10−4 shown in Fig. 5(a), in back-to-back operation, for the UWB sb#3 signal when compared with the UWB sb#1 (BER = 7 × 10−7) and sb#2 (BER = 2 × 10−5) signals. Additionally, the BER of 2×10−4 for the OFDM-BB signal in back-to-back operation, shown in Fig. 5(a), is partly due to the lower modulation index of mBB = 6.6% of this signal, which can be confirmed in Fig. 6(a).
The comparison of Figs. 5(a) and 5(b) shows also that noticeable BER degradations occur in OFDM-UWB sb#1–sb#3 with the increase of SSMF length when the signals are transmitted in coexistence (Fig. 5(a)) when compared to singly transmission (Fig. 5(b)). As an example, the OFDM-UWB sb#1 signal has a BER degradation from 2 × 10−7 (Fig. 5(b)) to 2 × 10−5 (Fig. 5(a)) when 100km of SSMF is considered. On the other hand, for the same SSMF length, the BER of the OFDM-BB signal has only a marginal degradation (from 4 × 10−4 to 5 × 10−4).
In summary, three important outcomes may be stated from these results:
- The intermodulation distortion induced by OFDM-UWB signals on OFDM-BB signal is very reduced;
- The intermodulation distortion of OFDM-UWB signals is very significant and is induced by OFDM-UWB sub-bands and OFDM-BB signals;
- This intermodulation distortion increases with the SSMF length. As a result, the main origin of intermodulation distortion is the SSMF dispersion-induced distortion.
The increase of the nonlinear distortion of OFDM-UWB signals with SSMF length would predict a monotonic BER degradation of OFDM-UWB sub-bands with the increase of SSMF length. However, Fig. 5(a) shows that significant BER improvements of OFDM-UWB sb#1–sb#3 are observed when the SSMF length increases from around 50km to 100km. To explain this apparent contradiction, Figs. 7(a)– 7(e) compare the electrical spectra of the coexisting OFDM-(BB+UWB) signals, for mBB = 6.6% and mUWB = 13%, at PIN output (at D - Fig. 1) in back-to-back operation and after transmission along 25km, 50km, 75km and 100km of SSMF, respectively. Figs. 7(b) and 7(c) show that the OFDM-UWB sb#1–sb#3 signals spectra are remarkably weakened and distorted for SSMF lengths of 25km and 50km when compared with back-to-back operation (Fig. 7(a)). This spectra degradation leads to the significant BER increase shown in Fig. 5(a), for SSMF lengths of 25km and 50km, and is attributed to the loss of the small dip of the fiber IM transfer function that appears for these lengths and at low-medium frequencies (around the frequencies of the OFDM-UWB sb#1–sb#3). The occurrence of this dip was already reported in . Notice that small differences of BER are observed between Fig. 5(a) and Fig. 5(b) for these SSMF lengths, indicating that intermodulation distortion is reduced for these lengths.
For SSMF lengths of 75km and 100km, Figs. 7(d) and 7(e) show that the OFDM-UWB sb#1–sb#3 signals spectra are strengthened, when compared to SSMF lengths of 25km and 50km, and almost equalized. The effect of SSMF length on the OFDM-UWB sb#1–sb#3 signals spectra strengthening shown in Figs. 7(d)– 7(e) is attributed to the gain induced by the fiber IM transfer function at the OFDM-UWB sb#1–sb#3 frequencies when SSMF distances of 75km and 100km are considered. This effect was already reported in . As a result, that effect can partly overcome the impact of the intermodulation distortion on the BER of the OFDM-UWB signals for those SSMF distances, leading to the improvement and partial equalization of the BERs of OFDM-UWB sb#1–sb#3 for those SSMF distances when compared to the SSMF distances of 25km and 50km, as shown in Fig. 5(a). It should be stressed that the gain of the fiber IM transfer function mitigates the weakness of the OFDM-UWB sb#1–sb#3 signals spectra due to the reduced DML bandwidth, but not the signal spectrum fluctuation resulting from the fluctuation of IM response of the DML, as can be seen more clearly in the spectrum of OFDM-UWB sb#3 shown in Fig. 7(e). This spectrum fluctuation causes additional performance degradation of UWB sub-bands, particularly of UWB sb#3 when compared to sb#1 and sb#2, as discussed in section 3.1.
Figure 7 shows also that SSMF transmission has negligible impact on OFDM-BB signal spectrum at PIN output (at D - Fig. 1). This is in accordance with the weak dependence of the BER of OFDM-BB signal on the SSMF length, shown in Fig. 5(a).
Further analysis of Fig. 7 shows a similar noise power level for all SSMF distances. As a result, no significant increase of intensity noise spectrum level with the accumulated dispersion occurs. As the intensity noise spectrum level increases with the accumulated dispersion when the laser noise is dominant , we conclude that, under the conditions of our experiments, the laser noise impact on the system performance is by far exceeded by the added optical noise. This assumption is corroborated by the quasi-constant slope monotonic BER degradation with the added optical noise power shown in Fig. 4. This feature allows drawing the same conclusion for the electrical receiver noise. Figure 7 shows also that no relevant power fading due to SSMF dispersion is observed, for SSMF lengths of 25,50,75 and 100km given that no large dips are observed in all spectra, which have been indicated as the usual effect of power fading on electrical spectrum .
The transmission performance of coexisting OFDM-BB and three OFDM-UWB sub-bands signals along LR-PONs using DMLs has been experimentally assessed. Reaches up to 100km with BERs lower than 5 × 10−4 for all signals with OSNR of 25dB@0.1nm have been shown, after optimization of the modulation indexes of OFDM-BB and OFDM-UWB signals. Additionally, when the OSNR takes values as high as 30dB@0.1nm, further BER improvements can be achieved, which decrease with the increase of the OFDM-UWB sub-band frequency.
It has been experimentally shown that the optimum performance of coexisting OFDM-(BB+UWB) signals is mainly imposed by the OFDM-BB and the OFDM-UWB sb#3 signals, due to the crucial role played by the combined effect of two issues: the intermodulation distortion induced by SSMF dispersion on the OFDM-UWB signals by the OFDM-BB and OFDM-UWB signals, and the additional degradation of the OFDM-UWB signals with higher frequency due to the reduced IM bandwidth of the DML. In our experiment, a higher number of UWB sub-bands is mainly limited by: i) the reduced DML bandwidth; ii) the increased intermodulation distortion; and iii) power fading in case of a sufficiently high number of sub-bands.
Those results indicate that, if usual solutions based on binary signals were used for GbE provision in coexistence with OFDM-UWB signals, higher nonlinear distortion induced on the OFDM-UWB signals by the GbE signal was expected, due to the lower spectral efficiency and less confined spectrum of conventional binary signals when compared with the OFDM-BB signal. Therefore, these results suggest that OFDM signals are an advantageous solution for GbE provision in LR-PONs when in coexistence with high capacity wireless services.
The achievements presented in this paper show that DMLs can be a competitive solution for cost-effective LR-PONs where coexisting GbE and wireless services using OFDM-UWB signals are delivered. Moreover, they indicate also that OFDM seems to be a suitable modulation format for transmitting in coexistence GbE and wireless services in a band-efficient manner.
This work was partly supported by the European Union FP7-ICT-2009-4-249142 FIVER project and by the PTDC/EEA-TEL/104358/2008 TURBO project from Fundação para a Ciência e a Tecnologia of Portugal.
References and links
1. T. Pleviak and V. Sahin, Next Generation Telecommunications Networks, Services, and Management (JohnWiley, New Jersey, USA, 2010). [CrossRef]
2. J. Ulm and B. Weeks, “Next play evolution: beyond triple play & quad play,” in Proceedings of IEEE International Symposium on Consumer Electronics (ISCE′2007), (Dallas, USA, 2007), (DOI: [CrossRef] ).
3. R. Davey, B. Grossman, M. Rasztovits-Wiech, D. Payne, D. Nesset, A. Kelly, A. Rafel, S. Appathurai, and S. Yang, “Long-reach passive optical networks,” IEEE/OSA J. Lightwave Technol. 27, 273–291 (2009). [CrossRef]
4. J. Kani, F. Bougart, A. Cui, A. Rafel, M. Campbell, R. Davey, and S. Rodrigues, “Next-generation PON - part I: technology roadmap and general requirements,” IEEE Commun. Mag. 47, 43–49 (2009). [CrossRef]
6. S. M. Lee, S. Mun, M. Kim, and C. Lee, “Demonstration of a long-reach DWDM-PON for consolidation of metro and acess networks,” J. Lightwave Technol. 25, 271–276 (2007). [CrossRef]
7. R. Llorente, T. Alves, M. Morant, M. Beltran, J. Perez, A. Cartaxo, and J. Marti, “Ultra-wideband radio signals distribution in FTTH networks,” IEEE Photon. Technol. Lett. 20, 945–947 (2008). [CrossRef]
8. D. Qian, N. Cvijetic, Y. Huang, J. Yu, and T. Wang, “100 km long reach upstream 36 Gb/s-OFDMA-PON over a single wavelength source-free ONUs,” in Proceedings of 35th European Conference on Optical Communication (ECOC′09), (Vienna, Austria, 2009), Paper 8.5.1. [PubMed]
9. L. Chen, J. Yu, S. Wen, J. Lu, Z. Dong, M. Huang, and G. K. Chang, “Novel scheme for seamless integration of ROF with centralized lightwave OFDM-WDM-PON system,” J. Lightwave Technol. 27, 2786–2791 (2009). [CrossRef]
10. H. Song, B. Kim, and B. Mukherjee, “Long-reach optical access networks: a survey of research challenges,” IEEE Commun. Surveys 12, 112–123 (2010). [CrossRef]
11. D. Fonseca, J. Morgado, and A. Cartaxo, “Transmission of multi-band OFDM-UWB signals along NG-FTTH networks using directly modulated lasers,” in Optical Fibre Communication Conference, OSA Technical Digest Series (CD) (Optical Society of America, 2011), paper OWK2.
12. M. Huang, J. Yu, D. Qian, N. Cvijetic, and G. Chang, “Lightwave centralized WDM-OFDM-PON network employing cost-effective directly modulated laser,” in Optical Fibre Communication Conference, OSA Technical Digest Series (CD) (Optical Society of America, 2009), paper OMV5.
13. S. Chuang, G. Liu, and P. Kondratko, “High-speed low-chirp semiconductor lasers,” in Optical Fiber Telecommunications. A: Components and Subsystems, I. Kaminow, T. Li, and A. Willner, eds. (Academic Press, San Diego, USA, 2008), chap. 3, pp. 53–80.
14. C. Chow, F. Kuo, J. Shi, C. Yeh, Y. Wu, C. Wang, Y. Li, and C. Pan, “100 GHz ultrawideband (UWB) fiber-to-the-antenna (FTTA) system for in-building and in-home networks,” Optics Express 18, 473–478 (2010). [CrossRef] [PubMed]
15. High rate ultra wideband PHY and MAC standard (2007). European Computer Manufacturers Association International Std. ECMA-368.
16. L. Bjerkan, A. Røyset, L. Hafskjaer, and D. Myhre, “Measurement of laser parameters for simulation of high-speed fiberoptic systems,” J. Lightwave Technol. 14, 839–850 (1996). [CrossRef]
17. T. Alves and A. Cartaxo, “Extension of the exhaustive Gaussian approach for BER estimation in experimental direct-detection OFDM setups,” Microwave and Optic. Technol. Lett. 52, 2772–2775 (2010).
18. S. Hunziker, “Low-cost fiber optic links for cellular remote antenna feeding,” in Radio over Fiber Technologies for Mobile Communications Networks, H. Al-Raweshidy and S. Komaki, eds. (Artech House, Norwood, USA, 2002), chap. 3, pp. 105–182.
19. M. Sakib, B. Hraimel, X. Zhang, K. Wu, T. Liu, T. Xu, and Q. Nie, “Impact of laser relative intensity noise on a multiband OFDM ultrawideband wireless signal over fiber system,” J. Opt. Commun. Netw. 2, 841–847 (2010). [CrossRef]
20. A. Cartaxo, “Small-signal analysis for nonlinear and dispersive optical fibres, and its application to design of dispersion supported transmission systems with optical dispersion compensation,” IEE Proc. Optoelectron.– Pt. J 146, 213–222 (1999). [CrossRef]
21. J. Morgado and A. Cartaxo, “OFDM-UWB signal distribution over long-haul-reach PON using directly modulated lasers,” in Proceedings of 12th International Conference on Transparent Optical Networks (ICTON′2010), (Munich, Germany, 2010), Paper Th.A2.5. [PubMed]