High-Bit-Rate Dynamically Reconfigurable WDM ... - OSA Publishing

Urban et al.

VOL. 1, NO. 2 / JULY 2009 / J. OPT. COMMUN. NETW.

A143

High-Bit-Rate Dynamically Reconfigurable WDM–TDM Access Network P. J. Urban, B. Huiszoon, R. Roy, M. M. de Laat, F. M. Huijskens, E. J. Klein, G. D. Khoe, A. M. J. Koonen, and H. de Waardt

Abstract—The intensification of traffic in the access network requires the development of novel architectural solutions for a reconfigurable network topology and components based on optical technologies. We present a hybrid ring-shaped wavelength division multiplexing (WDM)–time division multiplexing (TDM) passive optical network (PON) that is capable of providing bandwidth on demand at high bit rates in a transparent and dynamic manner. Our costefficient and scalable network architecture is based on integratable components such as a wavelengthagile optical networking unit and a microringresonator-based remote node. An appropriately modified control layer is introduced to manage the network. We also discuss the implementation of optical codes instead of time slots to take the step toward optical code division multiplexing (OCDM) WDM PONs that relieve the network of strict time scheduling of traffic and ranging. Therefore, an additional reduction of complexity in network management, improvement of network scalability, and a guarantee of fully symmetric traffic are foreseen for every user. Finally, we show a scenario for smooth migration from existing PON solutions to our WDM–TDM PON architecture.

Manuscript received October 31, 2008; revised January 15, 2009; accepted March 6, 2009; published June 30, 2009 共Doc. ID 103517兲. P. J. Urban, F. M. Huijskens, G. D. Khoe, A. M. J. Koonen, and H. de Waardt, are with the COBRA Institute, Eindhoven University of Technology, P.O. Box 512, Eindhoven, 5600 MB, The Netherlands (e-mail: [email protected]). B. Huiszoon was formerly with the COBRA Institute and is now with Universidad Autónoma de Madrid, Calle Tomás y Valiente, Madrid, 28049, Spain. R. Roy is with Telecommunication Engineering, University of Twente, 8202 Hogekamp, P.O. Box 217, Enschede, 7500 AE, The Netherlands. M. M. de Laat is with Genexis BV, Lodewijkstraat 1a, Eindhoven, 5652 AC, The Netherlands. E. J. Klein is with XiO Photonics BV, P.O. Box 1254, Enschede, 7500 BG, The Netherlands. Digital Object Identifier 10.1364/JOCN.1.00A143

1943-0620/09/02A143-17/$15.00

Index Terms—Fiber optics communications; Network topology; Protection and restoration; Wavelength routing; Modulators; Optical switching devices.

I. INTRODUCTION

I

t has been widely discussed that to satisfy users’ bandwidth demands optical fiber is the only practical solution for an access network [1,2]. Fast-growing traffic loads force the effort toward not only nearfuture solutions but also toward future-proof networks that can last for the next 25 years [3]. Table I shows the bandwidth requirements for some sample services [4,5]. Besides a singular session for a specific service, a number of terminals can be online at the same time, connected to a single access link (e.g., transmission of multiple high-definition TV feeds to one home). Moreover, the traffic pattern changes, for instance, from the before-noon businesscentric file transfer and video conference to afternoon entertainment-centric video-on-demand and voiceover-IP communication. Therefore the location of the traffic congestion changes on a specific time-scale basis [6]. This bandwidth-hungry scenario created by both content providers and consumers stimulates the development of novel components and network architectures that should not only be capable of transmitting data at high bit rates but should also be cost efficient. The latter is a necessity to make them particularly attractive for system and service providers. The physical layer of such a network has to be capable of providing bandwidth on demand, and, since the destination of the traffic load may change in time, the provision of the bandwidth should be made reconfigurable. This requires wavelength-agile optical network units (ONUs), reconfigurable optical add–drop multiplexers (OADMs), and an appropriate protocol layer [2]. Current installations in the access domain do not include these features and need to be upgraded from point-to-point and TDM-based point-to-

© 2009 Optical Society of America

A144

J. OPT. COMMUN. NETW. / VOL. 1, NO. 2 / JULY 2009

TABLE I DOWNSTREAM BANDWIDTH REQUIREMENTS ACCESS NETWORKS

FOR

FUTURE

Service

Bandwidth Consumption

SDTV HDTV 3D SDTV 3D HDTV Basic HSI Gaming Multimedia surfing Video-conf. and learning Telecommuting Voice-over-IP

2 Mb/ s per channel 8 Mb/ s per channel 63 Mb/ s per channel 187 Mb/ s per channel 5 Mb/ s average 10 Mb/ s average 8 Mb/ s average 3 Mb/ s per session 4 Mb/ s average 110 Kb/ s

multipoint (PtMP) to WDM–TDM-based PtMP [1]. Adding dynamic wavelength reallocation to the latter distributes the traffic congestion over a number of wavelengths more efficiently, and thus reduces blocking probability. However, the constraints originating from the time slotting of the data from or to different users transmitted on the same wavelength still requires a complex control and management layer. Here, optical code division multiplexing (OCDM) may alleviate this issue. Optical codes enable users to asynchronously access the network via optical orthogonal codes, which reduces the complexity, for example, because strict time scheduling and buffering is alleviated in the network [7]. Moreover, they facilitate network scalability because optical code division multiple access (OCDMA) has soft capacity properties; that is, no hard user limit is present that comes at the expense of multiple user interference. The other characteristics of a future-proof network, such as reliability, survivability, and coverage, can be improved by a network topology that is scalable and enables traffic routing. The cost efficiency of the network is enhanced by the integration of optical functions in photonic chips, which enables mass production to eventually provide a lower cost per device [8,9]. The most recent decade has brought numerous optical access research projects [1,10], and an evolutionary trend toward hybrid WDM–TDM PON can be noticed. Different architectures have been considered, such as splitter-based passive optical networks (PONs), (cascaded) arrayed waveguide grating (AWG)based PONs, amplified PONs, and PONs based on different wavelength spacings [11,12] as well as different topologies such as bus-and-tree or ring [13–15] and integrated metro-access architectures, for example [16]. In those efforts only fixed wavelength assignment has been considered. However, as access networks tend to connect a higher number of users [17,18], this growing amount of traffic should be dynamically managed, which brings the strong need for reconfigurability in the access domain [19–22].

Urban et al.

In this paper, we discuss in detail a novel architecture of a wavelength-flexible network, which can satisfy the requirements for future-proof access technology. Essentially, this design is studied in the Broadband Photonics project (BBPhotonics) [23] and comprises colorless integrated ONUs based on a reflective intensity modulator [24,25] and a fully reconfigurable integrated OADM based on microring resonators [26]. In order to support the concept of the BBPhotonics architecture, results showing the principle of operation of each component are described together with the overview of the system performance as analyzed in detail in [27,28]. In this paper, we also discuss the network management layer, which treats the architecture as a stack of PONs with the capability of switching between these PONs [29–31]. OCDM as a step toward the release from strict time scheduling of the traffic is proposed. Finally, a scenario is given for migration from a point-to-point (PtP) topology to the topology presented in this work. According to our best knowledge microring resonator technology has not been proposed as an accessdomain solution so far. The integrated colorless ONU designs for symmetric high-bit-rate operation up to 10 Gbit/ s are also key novel contributions of this paper. The OCDM technique has not yet been suggested in a wavelength-reconfigurable environment, and a reflective tunable ONU is introduced to handle the optical codes (OCs). The rest of this paper is organized as follows. Section II justifies the implementation of reconfigurability in the network, after which Section III presents the BBPhotonics ring-shaped network architecture in detail. Section IV then shows how control and management functions are realized in such a network. Two-dimensional OCs are enabled via the upgrades to the network presented in Section V. Finally, a migration scenario is discussed in Section VI, and the work is summarized in Section VII.

II. RECONFIGURABILITY The expected mix of services requested by a single ONU is presented in Table I. In a short-term evolution (including multiple HDTV channels, excluding 3DTV) a demand of average 60 Mbits/ s per ONU is expected. The traffic congestion caused mainly by video content transmission can be reduced by routing wavelength carriers from areas with lower bandwidth requirements to areas with higher bandwidth requirements. As a result the number of ONUs sharing the same wavelength channel is adapted dynamically. Such routing may be done by, e.g., adjustable wavelength-multicasting routers, with which settings

Urban et al.


Fig. 1. (Color online) System blocking probability versus relative system load for static and dynamic capacity allocation.

are optimized and adjusted to current bandwidth demands through a management protocol. The advantage of flexible capacity reallocation can be demonstrated by analyzing the call blocking probability. Given the actual traffic loads on all the wavelength channels, the call request of an ONU may not fit into its default wavelength channel, but may fit into another wavelength channel that still has sufficient capacity available. In the static wavelength assignment case this call of an ONU would have been blocked, whereas by using the flexible wavelength assignment it can be accepted. Therefore, the ondemand bandwidth allocation will decrease the call blocking probability remarkably, which implies that more calls can be accepted. Figure 1 shows the system blocking probability versus relative system load. The example system comprises eight wavelength channels with a capacity of 1.25 Gbits/ s each, and 256 ONUs, which generate Poisson-distributed calls with a data rate of 63 Mbits/ s or 125 Mbits/ s. The length of a call is assumed to be exponentially distributed. This may be a realistic model for exchanging files through the network. The normalized loading factor of the network is

Fig. 2.

A145

then defined as the average traffic generated by all the users (the average number of active users times their call bandwidth) divided by the aggregate capacity of all wavelength channels. Taking these assumptions, the number of users active at any given moment follows a discrete binomial distribution. Using the Chernoff ’s upper bound approximation, the system blocking probability versus relative system load has been calculated, under the assumption that granularity effects are negligible, which is a reasonable assumption when the call bandwidth is much smaller than the wavelength channel bandwidth [32]. It can be seen that the flexible capacity allocation reduces the blocking probability with respect to the fixed capacity allocation. For instance, for a blocking probability of 10−3, the system load may be doubled for 63 Mbit/ s calls, and more than doubled for 125 Mbit/ s calls. The above consideration shows that a higher loading factor for a given blocking probability is allowed in the wavelength-flexible network, as the resources at the optical line termination (OLT) are exploited more efficiently. This implies that the network operator has to install less equipment in the OLT, which results in cost reduction. Moreover, for a given amount of OLT equipment the network operator can accept more calls, and thus increase the revenues.

III. BBPHOTONICS NETWORK ARCHITECTURE The BBPhotonics network is a dynamically reconfigurable novel solution for the access domain. It is designed to connect 16 ONUs to a single remote rode (RN). There are four RNs connected in a ring topology to a central office (CO) by a standard single mode fiber (SSMF). The RN–RN and RN–ONU fibers are 1 km long and the CO is 20 km from the edge RNs as shown in Fig. 2. The physical topology does not determine

(Color online) BBPhotonics network architecture.

A146


Urban et al.

Such architecture allows the network to be extended with another RN connecting another 16 ONUs without interrupting the operation of the neighboring RN. In the current design 64 ONUs are served by 8 wavelength channel pairs. This provides the end user with the available bandwidth of 300 Mbits/ s to 1.25 Gbits/ s depending on whether a given wavelength pair serves a gigabit-capable PON (GPON) or a 10GPON system. Extending the network with additional RNs brings the need for more wavelength channels (a higher transmitter-to-receiver ratio in the CO). Currently, work is being performed toward the improvement of the splitting capability of the RN and thus more effective bandwidth utilization.

ADD/DROP PORTS

whether the downstream or upstream should go along the upper or lower branch of the ring, since bidirectional transmission over a single fiber is provided via bidirectional RNs. The optical switch connected to the circulator and to the two branches of the ring allows the communication to be directed via the upper or lower part of the ring; see Fig. 2.

COMMON INPUT

THROUGH PORT

2x2 SWITCH

RING

Fig. 3.

RING

(Color online) Eight-port OADM.

A. Central Office The CO contains a set of transmitters generating CW carriers and amplitude-modulated data signals for downstream transmission and a set of receivers for upstream termination, as shown in Fig. 2. Two AWGs are used as a WDM multiplexer and a WDM demultiplexer for downstream and upstream, respectively. The downstream and upstream traffic transmitted over a single fiber is split by a circulator. No direct communication between ONUs has been foreseen in this network, which means that all traffic is terminated at the CO. However, envisioned networking concepts may benefit from optical transparency in the access network [33]. A semipassive switch at the CO is added for protection and restoration purposes. When the switch detects a power loss in, for example, the upper branch of the ring, it switches from the cross state to the bar state such that transmission via the lower branch is enabled. Therefore, a break of the ring fiber causes a connection loss during the switching operation. Also, any maintenance or inserting a new RN will not seriously disturb network operation. CoarseWDM multiplexers are used to provide a control channel that is situated out of band with respect to the data signal channels in order not to interfere with or depend on them.

The OADM is equipped with thermally tunable microring resonators [26] as shown in Fig. 3 for an eightport device. The temperature dependency of the refractive index is used to apply a phase shift to the optical field. The thermal-optic effect is a slow process (milliseconds); thus it is suitable only for switching applications. With this device a single wavelength channel can be dropped to multiple users, or a single user can be assigned a wavelength at any given time. Figure 4 schematically illustrates the operation of a single microring resonator. The ring is connected to two waveguides in a four-port configuration (two inputs and two outputs). Consider a broadband input (multiple wavelength channels) at port 1. When the ring is in resonance for drop, the wavelength is dropped on port 4 together with all wavelengths that are separated by an integral multiple of the free spectral range (FSR) of the microring resonator. The re-

B. Remote Node The RN consists of an OADM, bidirectional optical amplification stage and coarse WDM multiplexers for control channel detection and transmission at each side as schematically shown in Fig. 2.

Fig. 4.

(Color online) Basic operation of a microring resonator.

Urban et al.

VOL. 1, NO. 2 / JULY 2009 / J. OPT. COMMUN. NETW. 1.40

THROUGH PORT

Fig. 5. (Color online) Prototype OADM with four microring resonators.

maining (nonresonant) wavelengths are transferred to port 2. The waveguides are situated orthogonally to enable the microring resonators to be placed in a matrix array as indicated in Fig. 3. The microring resonator structure is unidirectional in terms of common input to through-port transmission; therefore, in order to maintain the bidirectional traffic in the ring fiber, a 2 ⫻ 2 switch is necessary to keep optical signals running in the same direction through the OADM. If the traffic on the fiber ring changes direction, because of protection switching in the CO, this switch changes from the bar state to the cross state or vice versa. Preliminary measurements were done on a prototype OADM, which consists of a single row of four rings as shown in Fig. 5. A broadband source is connected to the common-input port, and the optical spectrum is measured by using an optical spectrum analyzer. Figure 6 shows the insertion loss at the through port of the OADM. Such a high insertion loss is a result of fiber-to-chip coupling losses, whereas the waveguide loss is around 3 dB/ cm. The microring resonators are designed to have a 500 GHz FSR and are tuned to the standard International Telecommunication Union (ITU) 50 GHz grid. The device shows a contrast ratio of about 18 dB with a smooth passband of 0.4 nm at −10 dB and 1.6 nm at −20 dB. Additionally, Fig. 7 shows the wavelength shift versus the applied voltage at the OADM. These results present a uniform response for the four different rings in the device shown in Fig. 5.

Wavelength shift [nm]

ADD/DROP PORTS COMMON INPUT

A147

1.20 1.00 Drop 1 Drop 2 Drop 3 Drop 4

0.80 0.60 0.40 0.20 0.00 0.0

1.0

2.0

3.0

4.0

5.0

Voltage [V]

Fig. 7. Resonant wavelength shift versus voltage applied to the heaters.

plifier or in a more cost-efficient manner by two separate semiconductor optical amplifiers combined with two circulators. When the network is enhanced with additional RNs the important factor for scalability appears to be the accumulated amplified spontaneous emission noise due to a cascade of the amplifiers. This has been investigated and discussed in [34]. C. Optical Network Unit The ONU contains a Mach–Zehnder (MZ) duplexer, which demultiplexes a modulated signal and a CW signal at its two outputs. As is shown in Fig. 8(a), a photodetector is connected to the lower output and a reflective semiconductor optical amplifier (RSOA) is connected to the upper output of the MZ duplexer. The CW signal is amplified and intensity modulated in the RSOA. It is reflected at the end facet of the RSOA and sent back to the CO with upstream data. The capability to provide gain and modulation at the same time reduces the need for additional amplification, and the wide amplification bandwidth of the

The bidirectional amplification can be provided either by a single bidirectional erbium-doped fiber am-

Fig. 6.

Through-port response of a prototype OADM.

Fig. 8. Solutions for the BBPhotonics ONU, including (a) RSOA, (b) REAM, and (c) MIM.

A148


Urban et al. P (a.u.)

upstream CW

downstream modulated

...

Fig. 9.

Eye diagrams of the reflective modulators shown in Fig. 8.

Bdrop

(a)

RSOA provides wavelength independency at the ONU. However, its electrical bandwidth is low because it is limited by the carrier lifetime [27]. Other solutions for integratable wavelengthagnostic reflective modulators are studied to overcome such a limitation, namely, a reflective electroabsorption modulator [REAM, Figure 8(b)] and a Michelsoninterferometer modulator (MIM) as shown in Fig. 8(c). The REAM works on the basis of a voltagecontrolled change in light absorption. Such single-port structures have been monolithically integrated with a semiconductor optical amplifier (SOA) [35,36]. These devices can operate at 10 Gbits/ s and have a large spectral range, up to 40 nm. Besides its very high modulation ability, the REAM reveals a very good compromise between driving voltage and high extinction ratio. The MIM works on the basis of phase-modulationto-amplitude-modulation conversion (like the MZ modulator) with a single 1:2 coupler and reflective coating at the end of each output waveguide [37]. The incident light is split equally into the two branches of the interferometer at the 1:2 coupler. Both signals travel through the waveguides to the mirror and are reflected back to the junction, where they recombine. By applying a differential voltage to the electrodes deposited on top of the two waveguides, a difference of refractive index in the two branches is created, as in the case of the MZ modulator. This causes a phase difference between the two reflected light signals, and, consequently, intensity modulation is realized. Figure 9 shows examples of simulated [38] and measured eye diagrams of the different modulators in a reflecting configuration. These measurements were performed with commercially available bulk components in a back-to-back setup (the REAM setup contained a traveling wave electroabsorption modulator and a circulator). It is clearly visible that for the RSOA the bit slopes are less steep than for the other modulators. Because of these high rise and fall times, which are a characteristic of the SOA, the extinction ratio is reduced for higher RF signal frequencies, and therefore such a modulator cannot provide high-speed modulation. The REAM and MIM show open and symmetric eye diagrams for bit rates much higher than the maximum allowable bit rate of the RSOA.

... f (THz)

FSRroadm

P (a.u.) Out1

Out2

f (THz)

(b)

2*FSRroadm

Fig. 10. (Color online) Wavelength architecture (a) at the OADM and (b) at the MZ duplexer outputs in the ONU.

D. Wavelength Panel The wavelength panel has to be matched to the FSR of the microring resonators in the OADM and to the periodicity of a MZ duplexer in the ONU. It also has to correspond to the ITU standard wavelength grid such that commercially available equipment may be employed. For this network the 1550 nm band is used, and the adjacent channels are spaced with 50 GHz. The channels are grouped in two bands, namely, downstream and upstream, as shown in Fig. 10(a), where the downstream band contains modulated wavelength channels and the upstream band contains CW carriers for remote modulation at the ONU. An upstream and a downstream channel, which are to be dropped to the same ONU, are spaced by a single FSR of the OADM (here 500 GHz). As shown in Fig. 10(b), the FSR of the MZ duplexer has to be twice the FSR of the OADM, that is, 1 THz, in order to separate the two channels. E. BBPhotonics System Performance A symmetrical 1.25 Gbit/ s transmission link from the CO via an RN to an ONU and backwards over a single fiber has been constructed (Fig. 11). The CO consists of two lasers generating optical carriers for downstream and upstream data transmission. Optionally (not included in Fig. 11), the channels are phase modulated in the CO in order to combat the destructive influence of backscattering by means of spectral broadening [39,40]. The receiver part of the CO consists of an AWG, an attenuator for bit error rate versus received optical power (ROP) measurement, a power tap to measure received optical power, and a commercial 1G receiver. The RN is composed of an

Urban et al.


Fig. 11.

(Color online) Setup for transmission experiment [28].

OADM with four add–drop ports, which is driven by a personal computer (PC). Because this OADM is a prototype device, it reveals substantial insertion loss 共15 dB兲 and low return loss 共17 dB兲. These significantly limit the performance of the system because of the low signal-to-crosstalk ratio of the received signal. To block the reflected power, two identical OADMs were applied together with two circulators to provide adding and dropping operations separately. The ONU is constructed with discrete elements, and it consists of a splitter, a tunable optical bandpass filter, an attenuator, a power tap, a commercial 1G receiver, a fixed bandpass filter, and a multiple-quantum-well RSOA [41] as shown in Fig. 11. The most important results achieved in these measurements are depicted in Fig. 12. The back-to-back

Fig. 12.

A149

(BtB) measurement of the downstream transmission (no fiber spans) shows a minor power penalty, and after the inclusion of 25 km of feeder fiber and 1 km of distribution fiber the results do not change. The backto-back upstream transmission measurement brings an around 2.5 dB power penalty, which is due to a combination of the lower Q factor, electrical bandwidth, and extinction ratio of the RSOA with respect to the MZ modulator used in the reference measurement. Inserting the fiber span causes an error floor, and the required bit error rate cannot be reached. This is due to the Rayleigh backscattering and Brillouin backscattering toward the ONU and toward the OLT. This influence is completely eliminated after applying phase modulation at the laser and enhancing its linewidth to ⌬␭3 dB = 50 pm.

Transmission experiment results. BER, bit error rate; ROP, received optical power.

A150


Urban et al.

Another important system experiment concerns the switching time of the OADM. In this measurement the CW signals are inserted by using a circulator after the OADM to avoid backreflections induced by poor return loss as mentioned earlier in this section (Fig. 13). Both CW signals are modulated simultaneously in the RSOA with the data generated by personal computer 1 (PC1). Only one channel passes through the OADM, which is controlled by PC2. Media converters are used for optoelectrical conversion. The switching time, designated as the time to achieve maximum throughput again when a microring has to be tuned from channel A to channel B spaced at a distance of 50 GHz, is measured by PC3 at 200 Mbit/ s by using user datagram protocol packets of 1470 bytes (Fig. 14). The switching time is determined to be approximately 6 ms, which will inherently lead to packet loss. However, when considering a time-critical service like voice communication, it will not disturb or disconnect the conversation. The obtained packet loss consists of a maximum 20 ms of voice data, which will not be noticed by the user. For other applications, retransmission of the packets will take place after time out. Detailed discussion of the experiments summarized in this section is given in [27,28].

IV. CONTROL AND MANAGEMENT The BBPhotonics network is considered a stack of quasi-independent logical PONs. The concept of bandwidth reallocation is shown in Fig. 15. The network from head end (HE) to customer premises equipment (CPE) is depicted as a two-stage switch. The firststage switching is done by a gigabit ethernet switch, which can route traffic to and from every OLT from

Fig. 13.

(Color online) Setup for switching experiment.

Fig. 14.

(Color online) Switching experiment results.

any of the ports toward the wide-area network (WAN) interface. The second-stage switch is the reconfigurable network itself, which can associate any ONU to any OLT based on the wavelength dropped toward it. Figure 15 shows two OLTs that are operating on a unique wavelength pair. Each OLT and the associated ONUs form a logical PON, named the red PON and the blue PON. The nominal bandwidth available to an ONU depends on the number of ONUs supported by each logical PON. If the number of ONUs supported by a logical PON is increased, then the nominal bandwidth available per ONU is reduced, while it increases when the number of ONUs is decreased. This obviously is an effect of using time slots in the network. The increase in the nominal bandwidth available to ONU1 in the blue PON is achieved by changing the wavelength assigned to ONU5, as shown in Fig. 15. The network reconfiguration and hence the consequent reallocation of the ONUs, is done by linear programming techniques to optimize the bandwidth distribution in the network [31]. A master controller (MC) is used, which monitors the network load and sets a configuration that is optimal for bandwidth availability to the end user. The MC communicates with the gigabit Ethernet switch and OLTs at the CO and with the node elements at the RNs. The RNs have local controllers (LC), which act as slave devices to the MC and do status monitoring and reconfiguration. The LCs are based on microcontrollers or embedded microcontrollers. All control and management for the network is done on an out-of-band communication channel. This channel works on 1310/ 1490 nm optics based on a 100Base-X communication link between the MC at the CO and LCs at the RNs. A two-fiber diversity in the link connection between the CO and every RN ensures fail-safe communication for up to a single link failure [29]. This out of band communica-

Urban et al.


Head End (HE)

Port 1 Port 2

Reconfig. Access Network

Customer Premises Equipment (CPE)

ONU 1

OLT 1

ONU 2

WAN Interface

ONU 3 ONU 4

OLT 2

Port u

A151

ONU 5

Port 1 Port 2

User Port 3 Interface Port 4 Port 5

2X5 switch

uX2 switch

Logical “Blue” PON Logical “Red” PON

Port 1 Port 2

ONU 1

OLT 1

ONU 2

WAN Interface

ONU 3 ONU 4

OLT 2

Port u uX2 switch Fig. 15.

ONU 5

Port 1 Port 2

User Port 3 Interface Port 4 Port 5

2X5 switch

(Color online) Concept of bandwidth reallocation.

tion channel for the network ensures independence in operation of EPON or any other standard MAC protocol for such a network. V. OPTICAL CODE-DIVISION MULTIPLEXING In this section the implementation of OCs is shown in the BBPhotonics network as a method to lower the complexity of the network management. First, the main characteristics of OCs are introduced, after which 2D coding is taken as a key technology. Then, a BBPhotonics network architecture is presented in which 2D OCs are implemented. The reader should note that the overall design is based on a system demonstrated in the literature.

A. Optical Code-Division Multiple Access Let us consider the traffic handling in the BBPhotonics network, where downstream traffic is centrally controlled at the CO while medium access has to be provided in case of upstream traffic. As explained in the previous sections, shared access in BBPhotonics is

enabled via time-division multiple access (TDMA) on WDM. Instead, OCs are proposed, which rely on communication via a unique and (pseudo-)orthogonal code. In an OC transmission system, the unique OC designates a single bit or each data bit, depending on the modulation scheme. The orthogonality of the code allows the carrier to be asynchronously shared with other users on the network. OCDMA has its primary application in the access network because it offers cost-effective network deployment and management combined with physical layer security [42]. In this case, deploying OCDMA on BBPhotonics enables broadcasting the data in the downstream direction. Moreover, on top of WDM it allows code reuse per wavelength channel. It is clear that OCDMA offers potential benefits to the reconfigurable BBPhotonics architecture, as it relieves the network of complex timemanagement schemes while offering dynamic networking behavior, inherently introducing transmission security and enabling fully symmetric communication channels; that is, bandwidth is equally available for upstream and downstream traffic. A dynamic assignment of different codes may be used to obtain a variable quality-of-service level per user.

A152


B. Two-Dimensional Incoherent Optical Coding Many different OCDMA systems have been demonstrated, which may be classified by their coding principle, coding domain, optical sources, and encoder– decoder implementation [43]. Taking a migration scenario into account, the aim in this work is to implement an optical coding technique that has the least impact on the BBPhotonics architecture while providing a solid performance. As a result a 2D time– wavelength incoherent OCDMA system is considered with fully integratable encoder–decoders [44]. A fouruser scenario has been experimentally demonstrated for such a system with a 2.5 Gbit/ s channel per user employing multiwavelength optical pulses, which are matched to the 50 GHz ITU grid. It has been shown that for similar systems the number of simultaneous users can be significantly increased by a number of techniques [45]. From here on the implementation of [44] in BBPhotonics is discussed, because the bandwidth of active components in the network only allows for 2D codes constructed by four wavelengths. Essentially, the 2D time–wavelength OCs are constructed by short pulses, which are arranged in time at four different wavelengths according to a hopping sequence. At the transmitter side the binary source data is modulated via on–off keying; thus a 2D OC is transmitted to represent a logical one and no information is sent in the case of a logical zero. At the receiver side the opposite time-delay configuration is used to reconstruct a high-intensity data signal via the incoherent summation of the pulses. On the physical level, a multiwavelength source produces a short-pulse train, which is modulated by an external modulator. The pulses are then encoded, via the method described above, after which they are called chips. During the encoding process, first, the pulses are demultiplexed by a wavelength demultiplexer (DEMUX), and most of them are given a defined time delay greater than zero. Then, the timeshifted pulses are combined by a wavelength multiplexer (MUX). This three-step operation is performed in a single device called an encoder. The receiver subsequently decodes the received data stream by rearranging the optical pulses with the opposite time delays. Since the photodetector is wavelength agnostic, the incoherent summation of pulses at different wavelengths results in a high-intensity output (autocorrelation peak). Note that this only occurs for matched delays, and, of course, amplification is required at various stages of the setup in order to compensate for losses in the system. The codes have been designed in such a way that a nonmatched set of delays results in a low-intensity output (crosscorrelation peak). In other words if encoded data is received from other users, only the cross-correlation

Urban et al.

peak is detected. The cross-correlation peak should therefore be as close to zero as possible through rigorous code design. An optical thresholder can be used in combination with optical time gating before detection in the receiver to virtually eliminate the detrimental effect of other users and thus improve the performance. Optical thresholding is preferred in such systems over optical time gating because the latter requires synchronization on a chip level and therefore effectively removes the advantage that OCDMA has over TDMA [46]. C. Implementation of 2D OCDMA in BBPhotonics The system in [44] employs a mode-locked supercontinuum (SC) and spectral slicing to generate the multiwavelength pulses. Four 50 GHz thin film filters are used to slice the SC spectrum according to the ITU grid. The pulses have a width of 9.5 ps before they are launched into the encoder. The encoder operates with a maximum pulse width of 9.8 ps in order to fit 41 chips into a 400 ps bit slot of a single carrier-hopping prime code according to [47]. A fiber-based encoder is employed, using components such as passive splitters and fiber delay lines, and only a downstream scenario is evaluated by the authors of [47]. According to the system in [44], four wavelengths (a quadruple) are necessary to construct 2D codes per each upstream and downstream band in BBPhotonics. As a result, a total optical bandwidth of 2 ⫻ 4 ⫻ FSROADM = 4 THz is required for deployment. Wellknown predispersion or postdispersion compensation schemes may be required in order to compensate for relative wavelength drifts because of the large total optical bandwidth. If we consider at least one shared pair of quadruples per RN, then a minimum of 4共RNs兲 ⫻ 8共2 ⫻ 4兲 = 32 wavelengths need to be continuously available. This number increases if dynamic bandwidth provisioning is enabled in the OCDMA– WDM network such that more than one pair of quadruples is provided per RN. 1) Central Office: An array of multiwavelength sources is readily installed at the CO as shown in Fig. 2 for the upstream and downstream traffic; hence these available optical resources should be reused in order to meet the system requirements of [44]. As mentioned before, additional optical sources may be required, depending on the demands in the network. A revised architecture of the CO is shown in Fig. 16 in which communication via 2D OCs is enabled. Basically, all sources are operated in CW mode, after which a single external modulator produces the pulse train at all wavelengths simultaneously with pulses smaller than 10 ps. The MUX and DEMUX operations are done by using AWGs with a matching channel spacing. For the sake of simplicity, the upper

Urban et al.


A153

…

circulator AWG

encoding and modulation

…

encoding and modulation

… AWG

decoder

…

… Fig. 16.

decoder

…

detector array

…

downstream band

…

…

…

AWG

…

AWG

external modulator

…

multiwavelength source

upstream band

CO enabling 2D OCDMA in BBPhotonics.

halves of the pulsed wavelength channels are directly multiplexed toward the single output of the CO; therefore these channels constitute the unmodulated and noncoded upstream bands. The lower half of the channels are encoded and modulated for downstream data transmission. Alternatively, the multiwavelength array in Fig. 16 can be replaced by SC sources or individual mode-locked laser diodes, which would replace all components in the dashed box. The SC sources used in [44] provide pulses with a relatively flat spectrum of about 12 nm, which are spectrally sliced. As such only several SC sources are required with respect to the total optical bandwidth of the system. Another option may be the recently shown ultra-fast integrated mode-locked laser diodes fabricated in InP / InGaAsP [48]. These stable, compact, and smallfootprint pulse sources introduce a cost reduction regarding the optical hardware required at the transmitter. Additionally, compared with the system in [44], the encoder and decoder in Fig. 16 may be equipped with AWGs as MUX–DEMUX components instead of passive splitters and thin film filters. Large-port AWGs are well-known devices and can be easily combined with tunable delay lines on a single photonic chip with other optical functions.

tuting the upstream and downstream band are alternatively positioned in the optical spectrum and spaced by FSROADM. The use of OCs in the network prohibits an upstream wavelength band to be dropped at multiple ports of one or more OADMs because user data is asynchronously multiplexed in the upstream direction. In other words, upstream data from a first port may not be partly dropped at a second port because of significant interference. In the case of the original TDMA this was allowed because of the use of access control via complex time management schemes. As a result of this restriction, an N : M passive coupler is added to the RN in order to broadcast all dropped wavelengths to the M connected ONUs as shown in Fig. 17 for the revised RN architecture. As shown, all N outputs of the OADM are connected to the passive coupler to enable dynamic bandwidthon-demand provisioning by dropping multiple upstream wavelength bands. If only one port is operated,

N -port OADM

ring up + down

…

N :M passive coupler up + down

2) Remote Node: The OADM in the RN operates in a way similar to that described in Subsection III.D, so the periodicity of the OADM is used to drop all eight equally spaced wavelengths. The wavelengths consti-

ring

up

… M ONUs

Fig. 17.

RN enabling 2D OCDMA in BBPhotonics.

A154


Urban et al.

data control and management AWG

drop ber microring resonator

τ1 τ2

AWG

REAM

τ3 MZ duplexer

AWG

τ4 τ5

AWG

PD

τ6 Fig. 18.

Architecture of the optical networking unit enabling 2D OCDMA in BBPhotonics.

all M ONUs can asynchronously access the network via the single broadcasted four-wavelength upstream band. If two or more ports are operated, multiple wavelength bands are dropped, which can be accessed by all ONUs by using their OC. The multiplexed encoded data streams in the upstream direction are automatically blocked by all ports not tuned to the correct wavelength band.

pologies, where each node in a layer aggregates nodes from lower layers [49]. Such topology is also very favorable for green-field deployments where the biggest problem is the large required investment (500–1000 €) per termination [9]. A layered topology enables stepby-step investment that is scalable to satisfy future bandwidth demands without drastic changes in the outside plant.

3) Optical Networking Unit: As is mentioned in Subsection III.C, the ONU is equipped with reflecting modulators. A reflective 2D OCDMA ONU is shown in Fig. 18, which enables the system to be used in a reflecting configuration on top of a dynamic or reconfigurable WDM scheme. The filter pattern layout of the architecture is shown in Fig. 19.

There are different intermediate solutions for fiberto-the-home enrollment, such as fiber to the cabinet, fiber to the premises, or fiber to the building. For any of the above there is a variety of connectivity topologies, namely, PtP, TDM-PON, WDM-PON, or WDM– TDM-PON.

The principle of operation is as follows. The microring resonator has the same FSR and drop band as the OADM and therefore can tune to the same wavelength band. The periodicity of the tunable MZ duplexer equals 2 ⫻ FSROADM, and it is used to demultiplex the upstream and downstream channels at its two outputs. Both are controlled by a control and management block. The modulated and encoded downstream data is processed by a decoder, while the unmodulated and noncoded pulse train is processed by the reflective encoder. The reader should note that the time delays in the reflective encoder are half of the required time shift (␶1 = 0.5␶4, ␶2 = 0.5␶5, etc.) because after modulation by the REAM the pulses pass a second time through the encoder. The AWGs with a channel spacing of 2 ⫻ FSROADM ensure the correct demultiplexing of the pulses at the encoder.

The main advantage of PtP (home run) is that, apart from providing a complete channel with fully symmetric bandwidth per user, all maintenance is in two places only, that is, at the CO and at the ONU. However, a significant deployment cost is related to such installation mainly due to investment in the equipment at the CO. TDM-PON (either in distributed or cascaded splitter configuration) brings some alleviation in the initial investment, as some resources are shared among a number of users. Here, the drawback is the reduction of available bandwidth per user due to a high splitting ratio, which divides a single wavelength channel into user-dedicated time slots.

microring resonator upper output MZ duplexer up

down

up

AWG out

VI. MIGRATION SCENARIO Whichever technology is used to provide fiber to the home, a general conclusion can be drawn, namely, fiber to the home is extending the metro networks into the local loop, and this requires the reduction of traffic congestion in the access nodes. This can be achieved by creating a layered structure in existing network to-

FSRoadm Bdrop 2*FSRoadm

Fig. 19. Filter scheme applied in the the optical networking unit enabling 2D OCDMA in BBPhotonics.

Urban et al.


A155

WDM-PON provides more transmission channels and so the network stretches to cover a larger number of users. Moreover, such a network upgrade is not related to significant costs, as the fiber plant can be reused, as will be discussed below in this section. When considering an open-access network model, PtP topology enables sharing the infrastructure by a number of service providers simply by assigning a specific link or group of links to one operator within a given access network. Therefore, a PtP topology based on dedicated fiber (home run) or dedicated wavelength (WDM-PON) is advantageous over PtMP TDM-PON. Nowadays most of the green-field installations provide PtP connectivity as depicted in Fig. 20(a). Thus, when considering new services in the network a migration scenario to a future-proof network is needed. Such an upgrade should be as least distractive to connected users as possible (short down time) and should involve only minor investment. In order to avoid high cost related to the initial upgrade, a transition from PtP to several TDM-PON systems is proposed, which means that several passive splitters have to be installed somewhere between the CO and the ONUs at a central location. Depending on the topology it may be a cabinet in the street, building, etc. For this purpose a single fiber can be reused from the previous PtP installation to connect a splitter to the CO. The PtP links from the splitter to ONUs may reuse the existing fibers and ducts. In the electric domain the upgrade requires installation of all TDM protocol-related resources (i.e., control unit), which may become a significant investment at this stage. On the other hand, the number of optical sources at the CO is reduced to the number of feeder fibers as shown in Fig. 20(b). The step toward WDM-PON is to exchange the equipment in the splitting points from passive splitters to wavelength (de)multiplexers [14] and to implement an appropriate set of optical sources at the CO as well as wavelength-agile ONUs. The BBPhotonics project upgrades the access network to a hybrid WDM–TDM scheme. Therefore the passive splitters are changed to reconfigurable OADMs. As explained in the paper, each wavelength pair in a WDM–TDMPON serves a separate PON system (e.g., GPON). This requires as many OLTs as wavelength pairs. To dynamically reallocate the bandwidth over all PONs, the MC and LCs have to be implemented at the CO and RN, respectively, as explained in Section IV. In order to improve the network protection the CO–RN distribution fiber topology is changed to a ring-shaped topology. Together with the switch at the CO shown in Fig. 2, it provides a protection path in case of a fiber break in the feeder part. The connectivity at this stage does not change with respect to the

Fig. 20. (Color online) Migration scenario from (a) existing PtP topology via (b) PtMP topology to (c) hybrid ring–PtMP topology.

TDM-PON. If some of the feeder fibers pass by the CO as indicated in Fig. 20(c), the existing fiber plant can be reused, leading to lower upgrade costs. When OCDM is considered as a solution for TDM release, some more changes in the network are required, which were described in Section V. They in-

A156


volve the upgrade of the CO and ONUs, where the encoders and decoders have to be added. Additionally, a single high-speed modulator to produce the short pulses may be included in the CO, and that requires another stage of AWGs as given in Fig. 16. The RNs are upgraded with a passive splitter to which all drop ports are connected. The fiber plant does not require any change, as the existing installation may be used again, which limits the cost of network upgrade. Concluding and considering the above, it is clear that migration toward a high-bandwidth and reconfigurable fiber-optic access layer involves a great investment in the network; however, this will eventually be outweighed by the additional functionalities offered and therefore a return of investment.

VII. SUMMARY We showed the clear need for dynamic bandwidth allocation and presented a reconfigurable WDM– TDM-PON architecture that can deliver highbandwidth services on demand to residential users. The network is constructed by novel cost-efficient components. The OADM architecture is introduced in the access domain for the first time as well as the three solutions for the colorless ONU, all supported by relevant results. We also showed error-free transmission in a testbed that proved the feasibility concept of the BBPhotonics architecture. The control and management layer of a wavelength-flexible WDM–TDMPON has been discussed together with the implementation. We also proposed the replacement of TDM with 2D incoherent OCDM, which can reduce the management complexity of the network and provides symmetric traffic to and from every user. A novel design of a reflective ONU has been shown, for the first time to our knowledge, which enables us to use 2D incoherent OCs in a colorless ONU. Finally, we discussed the migration scenario from existing PtP and TDM-PON solutions via WDM– TDM-PON to a WDM–OCDM-PON network.

ACKNOWLEDGMENTS This work was done in the Freeband BBPhotonics project (http://bbphotonics.freeband.nl) with support of the COBRA SWOOSHING project. Freeband is sponsored by the Dutch Government under contract BSIK 03025, and the Netherlands Organisation for Scientific Research (NWO) is acknowledged for funding the COBRA Institute via the NRC Photonics grant.

Urban et al.

REFERENCES [1] C.-H. Lee, W. V. Sorin, and B. Y. Kim, “Fiber to the home using a PON infrastructure,” J. Lightwave Technol., vol. 24, pp. 4568–4583, 2006. [2] A. M. J. Koonen, “Fiber to the home/fiber to the premises: what, where, and when?” Proc. IEEE, vol. 94, no. 5, pp. 911– 934, May 2006. [3] M. J. O’Mahony, C. Politi, D. Klonidis, R. Nejabati, and D. Simeonidou, “Future optical networks,” J. Lightwave Technol., vol. 24, pp. 4684–4696, 2006. [4] R. W. Heron, T. Pfeiffer, D. T. van Veen, J. Smith, and S. S. Patel, “Technology innovations and architecture solutions for the next-generation optical access network,” Bell Syst. Tech. J., vol. 13, pp. 163–181, 2008. [5] Y. Xiao, X. Du, J. Zhang, F. Hu, and S. Guizani, “Internet protocol television (IPTV): the killer application for the nextgeneration internet,” IEEE Commun. Mag., vol. 45, no. 11, pp. 126–134, Nov. 2007. [6] K. Cho, K. Fukuda, H. Esaki, and A. Kato, “The impact and implications of the growth in residential user-to-user traffic,” in Proc. 2006 Conf. on Applications, Technologies, Architectures, and Protocols for Computer Communications, Pisa, Italy, Sept. 11–15, 2006, pp. 207–208. [7] B. Huiszoon, L. M. Augustin, E. A. J. M. Bente, H. de Waardt, G. D. Khoe, M. K. Smit, and A. M. J. Koonen, “Integrated Mach-Zehnder-based spectral amplitude OCDMA on a passive optical network,” IEEE J. Sel. Top. Quantum Electron., vol. 13, pp. 1487–1496, 2007. [8] R. Rubenstein, “Optical integration moves back onto the agenda,” Fibre Syst. Eur., pp. 14–17, May 2006, http:// fibresystems.org/cws/article/articles/24945. [9] E. Iannone, “Passive optical networks must evolve to survive,” Fibre Systems Eur., pp. 26–27, Jan. 2008, http:// fibresystems.org/cws/article/magazine/32042. [10] P. W. Shumate, “Fiber-to-the-home: 1977–2007,” J. Lightwave Technol., vol. 26, pp. 1093–1103, 2008. [11] K. Grobe and J.-P. Elbers, “PON in adolescence: from TDMA to WDM-PON,” IEEE Commun. Mag., vol. 46, no. 1, pp. 26–34, Jan. 2008. [12] A. Banerjee, Y. Park, F. Clarke, H. Song, S. Yang, G. Kramer, K. Kim, and B. Mukherjee, “Wavelength-division-multiplexed passive optical network (WDM-PON) technologies for broadband access: a review,” J. Opt. Netw., vol. 4, pp. 737–757, 2005. [13] G. Keiser, FTTX Concepts and Applications, Hoboken, NJ: Wiley, 2006. [14] F.-T. An, D. Gutierrez, K. S. Kim, J. W. Lee, and L. G. Kazovsky, “SUCCESS-HPON: a next-generation optical access architecture for smooth migration from TDM-PON to WDMPON,” IEEE Commun. Mag., vol. 43, no. 11, pp. S40–S47, Nov. 2005. [15] J. Prat, P. E. Balaguer, J. M. Gené, and O. Diaz, Fiber-to-theHome Technologies, Norwell, MA: Kluwer Academic, 2002. [16] G. Talli, C. W. Chow, P. Townsend, R. Davey, T. de Ridder, X.-Z. Qui, P. Ossieur, H.-G. Krimmel, D. Smith, I. Lealman, A. Poustie, S. Randel, and H. Rohde, “Integrated metro and access network PIEMAN,” in 12th European Conf. on Networks and Optical Communications—NOC 2007, Stockholm, Sweden, June 18–21, 2007, pp. 493–500. [17] I. van de Voorde, C. M. Martin, and J. Vandewege, “The SuperPON demonstrator: an exploration of possible evolution paths for optical access networks,” IEEE Commun. Mag., vol. 38, no. 2, pp. 74–82, Feb. 2000. [18] S. A. Jabar, “Alternative architectures for bidirectional single mode fiber SuperPON 512 ONU, 100 km,” in IEEE Region 10 TENCON 2005, Melbourne, Australia, Nov. 21–24, 2005, pp. 1–5. [19] A. M. J. Koonen, “Flexible wavelength multiplexing techniques

Urban et al.

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

for broadband fibre access networks,” Telektronikk, vol. 95, pp. 119–128, 1999. A. M. J. Koonen, T. Muys, C. van der Plaats, S. M. Heemstra de Groot, H. J. H. N. Kenter, I. G. M. M. Niemegeers, and F. N. C. Slothouber, “TOBASCO: an innovative approach for upgrading CATV fiber-coax networks for broadband interactive services,” IEEE Commun. Mag., vol. 35, no. 4, pp. 76–81, Apr. 1997. A. M. J. Koonen, K. Steenbergen, F. Janssen, and J. Wellen, “Flexibly reconfigurable fiber-wireless network using wavelength routing techniques: the ACTS project AC349 PRISMA,” Photonic Network Commun., vol. 3, pp. 297–306, 2001. E. Yetginer and E. Karasan, “Dynamic wavelength allocation in IP/WDM metro access networks,” IEEE J. Sel. Areas Commun., vol. 26, pp. 13–27, 2008. P. J. Urban, E. G. C. Pluk, A. M. J. Koonen, G. D. Khoe, and H. de Waardt, “First design of dynamically reconfigurable broadband photonic access networks (BBPhotonics),” in 2005 IEEE/ LEOS Symp. Benelux Chapter Proc., Mons, Belgium, Dec. 1–2, 2005, pp. 117–120. L. Xu, X. J. M. Leijtens, M. J. H. Sander-Jochem, T. de Vries, Y. S. Oei, P. J. van Veldhoven, R. Nötzel, and M. K. Smit, “InP based polarization insensitive tunable duplexer and integrated reflective transceiver,” in European Conf. on Integrated Optics, Copenhagen, Denmark, Apr. 25–27, 2007, paper WA2. P. J. Urban, E. J. Klein, L. Xu, E. G. C. Pluk, A. M. J. Koonen, G. D. Khoe, and H. de Waardt, “1.25– 10 Gbit/ s reconfigurable access network architecture,” in 9th Int. Conf. on Transparent Optical Networks, 2007. ICTON ’07, Rome, Italy, July 1–5, 2007, pp. 293–296. E. J. Klein, P. J. Urban, G. Sengo, L. T. H. Hilderink, M. Hoekman, R. Pellens, P. van Dijk, and A. Driessen, “Densely integrated microring resonator based photonic devices for use in access networks,” Opt. Express, vol. 15, pp. 10346–10355, 2007. P. J. Urban, E. J. Klein, L. Xu, E. G. C. Pluk, A. M. J. Koonen, G. D. Khoe, and H. de Waardt, “1.25– 10 Gbit/ s bidirectional link in an access network employing a reconfigurable optical add/drop multiplexer and a reflective semiconductor optical amplifier,” in 10th Anniversary Int. Conf. on Transparent Optical Networks, 2008. ICTON 2008, Athens, Greece, June 22– 26, 2008, vol. 4, pp. 166–169. P. J. Urban, E. G. C. Pluk, M. M. de Laat, F. M. Huijskens, G. D. Khoe, A. M. J. Koonen, and H. de Waardt, “1.25 Gbit/ s transmission over an access network link with tuneable OADM and a reflective SOA,” IEEE Photon. Technol. Lett., vol. 21, no. 6, pp. 380–382, 2009. R. Roy, G. Manhoudt, C. Roeloffzen, and W. van Etten, “Control and management scheme in a DWDM EPON,” in 2006 Int. Conf. on Transparent Optical Networks, Nottingham, UK, June 18–22, 2006, vol. 1, pp. 151–154. R. Roy, G. G. H. Roeloffzen, and W. van Etten, “Switched DWDM ethernet passive optical networks,” in 2006 Annu. Symp. of the IEEE/LEOS Benelux Chapter, Eindhoven, The Netherlands, Nov. 30–Dec. 1, 2006, pp. 169–172. R. Roy, G. Manhoudt, and W. van Etten, “Bandwidth redistribution techniques for extended EPON based multiwavelength networks,” in 9th Int. Conf. on Transparent Optical Networks, 2007. ICTON ’07, Rome, Italy, July 1–5, 2007, vol. 4, pp. 80–83. A. M. J. Koonen, A. Ng’oma, G.-J. Rijckenberg, M. Garcia Larrode, P. J. Urban, H. de Waardt, J. Yang, H. Yang, and H. P. A. van den Boom, “How deep should fiber go into the access network?” in 33rd European Conf. and Exhibition on Optical Communication—ECOC 2007, Berlin, Germany, Sept. 16–20, 2007, pp. 114–117. B. Huiszoon, F. T. H. den Hartog, M. Garcia Larrode, and A. M. J. Koonen, “Layer 2 and 3 contention resolution and radio-


[34]

[35]

[36]

[37]

[38] [39]

[40]

[41] [42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

A157

over-fiber in OCDMA PON for transparent optical access in personal networks,” J. Lightwave Technol., vol. 26, pp. 1752– 1764, 2008. P. J. Urban, E. G. C. Pluk, E. J. Klein, A. M. J. Koonen, G. D. Khoe, and H. de Waardt, “Simulation results of dynamically reconfigurable broadband photonic access networks (BBPhotonics),” in 2nd IET Int. Conf. on Access Technologies, Cambridge, UK, June 21–22, 2006, pp. 89–92. A. Garreau, J. Decobert, C. Kazmierski, M. C. Cuisin, J. G. Provost, H. Sillard, F. Blache, D. Carpentier, J. Landreau, and P. Chanclou, “10 Gbit/ s amplified reflective electroabsorption modulator for colorless access networks,” in 2006 Int. Conf. on Indium Phosphide and Related Materials Conf. Proc., Princeton, NJ, May 7–11, 2006, pp. 68–170. I. Tafur-Monroy, F. Ohman, K. Yvind, L. Jin-Christiansen, J. Mørk, C. Peucheret, and P. Jeppesen, “Monolithically integrated reflective SOA-EA carrier re-modulator for broadband access nodes,” Opt. Express, vol. 14, pp. 8060–8064, 2006. F. Zernike and J. C. Webster, “Amplitude modulators based on the Michelson interferometer,” IEEE Trans. Sonics Ultrason., vol. 3, pp. 163–167, 1972. VPI TransmissionMaker, 2008, http://www.vpiphotonics.com. P. J. Urban, A. M. J. Koonen, G. D. Khoe, and H. de Waardt, “Mitigation of reflection-induced crosstalk in a WDM access network,” in Optical Fiber Communication Conf. and Expo. and Nat. Fiber Optic Engineers Conf., San Diego, CA, Feb. 24– 28, 2008, OSA Technical Digest (CD), Washington, DC: Optical Society of America, 2008, paper OThT3. P. J. Urban, A. M. J. Koonen, G. D. Khoe, and H. de Waardt, “Coherent crosstalk-suppression in WDM access networks employing reflective semiconductor optical amplifiers,” in 33rd European Conf. and Exhibition on Optical Communication— ECOC 2007, Berlin, Germany, Sept. 16–20, 2007, pp. 642–643. CIP Technologies, 2008, http://www.ciphotonics.com. K. Fouli and M. Maier, “OCDMA and optical coding: principles, applications, and challenges,” IEEE Commun. Mag., vol. 45, no. 8, pp. 27–34, Aug. 2007. X. Wang, “Keys towards practical OCDMA networks,” in Proc. IEEE PHOTONICS 2004, Cochin, India, Dec. 8–11, 2004, paper NET2.2. V. Baby, I. Glesk, R. J. Runser, R. Fischer, Y.-K. Huang, C. -S. Bres, W. C. Kwong, T. H. Curtis, and P. R. Prucnal, “Experimental demonstration and scalability analysis of a four-node 102-Gchip/ s fast frequency-hopping time-spreading optical CDMA network,” IEEE Photon. Technol. Lett., vol. 17, pp. 253– 255, 2005. A. E. Wilner, P. Saghari, and V. R. Arbab, “Advanced techniques to increase the number of users and bit rate in OCDMA networks,” IEEE J. Sel. Top. Quantum Electron., vol. 13, pp. 1403–1414, 2007. K.-I. Kitayama, X. Wang, and N. Wada, “OCDMA over WDM PON—solution path to gigabit-symmetric FTTH,” J. Lightwave Technol., vol. 24, pp. 1654–1662, 2006. G.-C. Yang and W. C. Kwong, Prime Codes With Applications to CDMA Optical and Wireless Networks, Norwoord, MA: Artech House, 2002. M. J. R. Heck, “Ultrafast integrated semiconductor laser technology at 1.55 ␮m,” Ph.D. thesis, Eindhoven University of Technology, Eindhoven, The Netherlands, 2008. “Operators take on the FTTH challenge,” Fibre Syst. Eur., pp. 21–24, Nov. 2007, http://fibresystems.org/cws/article/magazine/ 32040.

A158


P. J. Urban (S’05) was born in Szczecin, Poland, in 1981. He received the M.Sc. degree in electrical engineering from the Szczecin University of Technology, Szczecin, Poland, in 2004, where he was also a student assistant from 2003 to 2005 in the Optical Telecommunication and Photonics Group. In 2004 he pursued doctoral studies at Szczecin University of Technology in the field of nonlinear optics. In 2005 he joined the COBRA Research Institute at the Eindhoven University of Technology, The Netherlands, and changed the subject of his Ph.D. research to WDM/TDM access networks. He is currently involved in research on next-generation broadband access network architectures involving network reconfigurability and bandwidth on-demand provision. He acts as a reviewer for the IET’s Electronics Letters and IEEE Photonics Technology Letters. Since 2005 he has been an IEEE student member and in 2006–2008 a member of the IEEE/LEOS Benelux Student Chapter Board. B. Huiszoon is a post-doctoral researcher at the Universidad Autonoma de Madrid (UAM) in Madrid, Spain. He received the M.Sc. and Ph.D. degrees in electrical engineering from the Eindhoven University of Technology (TU/e), Eindhoven, The Netherlands, in 2003 and 2008, respectively, and joined UAM in December 2008 with a Juan de la Cierva contract from the Spanish Ministry of Science and Innovation. He participated and contributed to a number of national- (RETINA, RGE, DIOR) and European- (e-Photon/ONe⫹, BONE, ISIS) funded research projects. He received the Valorisation Grant I in June 2007 from the Dutch Ministry of Economic Affairs, which supports technology transfer from academics to industry. His expertise is in the field of optical access networks and systems, OCDMA, hybrid optical–wireless networks, and personalized service delivery. The results of Dr. Huiszoon’s M.Sc. and Ph.D. theses were awarded with the TU/e Mignot Prize of 2004 (best thesis), the IEEE LEOS Graduate Student Fellowship of 2008, and the Dutch Veder Prize of 2008. He is member of the OCDMA TPC at the LEOS Summer Topicals 2009 and was on the founding board of the IEEE/ LEOS Benelux Student Chapter from November 2004 until May 2007. Dr. Huiszoon (co-)authored more than 30 papers in journals and conferences, holds a Dutch patent, and is a member of the IEEE (2004), OSA (2009), and NERG (2009). R. Roy was born in New Delhi, India. He received his M.Sc. in physics (1997) and M.Tech. in optoelectronics and optical communications (1999) from the Indian Institute of Technology (IIT) Delhi. He worked for Tejas Networks for four years developing hardware systems and subsystems for NG SDH/SONET products and designing DWDM networks and optical systems. He holds three patents related to OSNR improvements in low-cost DWDM systems. His current interests include design of high-speed transmission systems for use in metro and long-haul systems, network studies on photonic broadband access networks, and optimization techniques for use in reconfigurable networks. He is currently pursuing his Ph.D. in the Telecommunication Engineering Department of the University of Twente and working as a Senior Hardware Design Engineer in AimValley BV, Hilversum, The Netherlands.

Urban et al.

WDM/TDM schemes.

PON

M. M. de Laat was born in Eindhoven, The Netherlands, in April 1983. He received his M.Sc. degree in electrical engineering from the Electro-Optical Communication Systems Group of the Eindhoven University of Technology, The Netherlands, in 2008. Presently he is working as an Application Engineer at Genexis B.V. in Eindhoven. His research interests are in the area of nextgeneration optical access networks, including WDM PON systems and hybrid systems using bandwidth reconfiguration

F. M. Huijskens graduated in applied physics at the Technical College of Dordrecht, The Netherlands, in 1979. From 1981 to 1984, he was an Electronic Test Engineer at Siemens Gammasonics. In 1985, he joined the Electro-Optical Communications Group of Eindhoven University of Technology, Eindhoven, The Netherlands. His work involved research on passive fiber couplers and technical support on projects concerning phase and polarization diversity coherent systems. He contributed to the development of optical cross-connect demonstrators and to packaging technologies of optical integrated devices. Recently he has assisted in the fields of optical packet switching and characterization of photonic devices. E. J. Klein obtained his M.Sc. degree from the Electrical Engineering Department of the University of Twente in 2002. After receiving his Ph.D. at the Integrated Optical MicroSystems Group of the University of Twente, he continued working as a Post-Doc in the same group before joining LioniX BV in 2008, working as a Product Specialist, focusing on microring-resonator-based PLC devices for use in optical telecommunication. In the same year, he won the prestigious Dutch Veder award based on his research on microring resonators. He is currently working for XiO Photonics as a Design Engineer. Dr. Klein has (co)authored more than 35 publications in journals and conferences. His primary interest is the development of complex photonic devices in TriPleX waveguide technology. G. D. Khoe received the degree of Elektrotechnisch Ingenieur, cum laude, from the Eindhoven University of Technology, Eindhoven, The Netherlands, in 1971. He decided to start research at the Dutch Foundation for Fundamental Research on Matter (FOM) Laboratory on Plasma Physics, Rijnhuizen. In 1973 he moved to the Philips Research Laboratories to start research in the area of optical fiber communication systems. In 1983, he was appointed as Part-Time Professor at Eindhoven University of Technology. He became a Full Professor at the same university in 1994 and is currently Chairman of the Department of Telecommunication Technology and Electromagnetics (TTE). Most of his work has been devoted to single-mode fiber systems and components. Currently his research programs are centered on ultrafast all-optical signal processing, high-capacity transport systems, and systems in the environment of the users. He has more than 40 U.S. patents and has authored and co-authored more than 150 papers, invited papers, and chapters in books. His professional activities include many conferences, where he has served on technical committees, management committees, and advisory committees as a member or chairman. He has been involved in journal activities, as associate editor, as a member of the advisory board, or as a reviewer. In Europe, he is closely involved in Research Programs of the European Community and in Dutch national research programs, as a participant, evaluator, auditor, program committee member, and member of the steering committee. He is one of the founders of the Dutch COBRA University Research Institute and one of the three recipients of the prestigious Top Research Institute Photonics grant that is awarded to COBRA in 1998 by the Netherlands Ministry of Education, Cul-

Urban et al. ture and Science. In 2001, he brought four European groups together to start a new international alliance called the European Institute on Telecommunication Technologies (eiTT). He has served in the IEEE/LEOS organization as the European Representative in the BoG, VP Finance & Administration, VP Membership, BoG Elected Member, President, and member of the Executive Committee of the IEEE Benelux Section. He was the founder of the LEOS Benelux Chapter. He has been an IEEE Fellow since 1991, an OSA Fellow since 2006, and received the MOC/GRIN Award in 1997. A. M. J. Koonen (F’07-SM’01-M’00) received his M.Sc. (cum laude) in electrical engineering from Eindhoven University of Technology in 1979. He spent more than 20 years at Bell Laboratories in Lucent Technologies as a Technical Manager of applied research. He has also been a Bell Labs Fellow since 1998 (the first one in Europe). Next to his industrial position, he has been a Part-Time Professor at Twente University from 1991 to 2000. Since 2001, he has been a Full Professor at Eindhoven University of Technology in the Electro-Optical Communication Systems Group, which is a partner in the COBRA Institute; since 2004, he has been the Chairman of this group. His main interests are currently in broadband fiber access networks and in optical packet-switched networks. He has initiated and led several European and national R&D projects in this area, on label controlled optical packet routed networks (the EC FP5 IST project STOLAS), dynamically reconfigurable hybrid fiber access networks (EC FP4 ACTS TOBASCO on fiber-coax, EC FP4


A159

ACTS PRISMA on fiber-wireless, EC FP5 IST HARMONICS on packet-switched access), and short-range multimode (polymer) optical fiber networks. Presently, he is involved in a number of access/ in-home projects in the Dutch Freeband program, in the Dutch IOP Generieke Communicatie program, and in the EC FP6 IST Broadband for All program (MUSE, e-Photon/ONe⫹, POF-ALL). H. de Waardt was born in Voorburg, The Netherlands, in December, 1953. He received the M.Sc.E.E. and the Ph.D. degrees from the Delft University of Technology, The Netherlands, in 1980 and 1995, respectively. In 1981, he started his professional career in the Physics Department at PTT Research in Leidschendam, The Netherlands, where he worked on the performance issues of optoelectronic devices. In 1989 he moved to the Transmission Department and became involved in WDM high-bit-rate optical transmission. In 1995 he was appointed as an Associated Professor at the Eindhoven University of Technology (TU/e), Eindhoven, The Netherlands, in the area of high-capacity trunk transmission. He coordinated the participation of TU/e in ACTS Upgrade, ACTS BLISS, ACTS APEX, and IST FASHION. Presently he serves as the Project Leader of the national research initiative Freeband Broadband Photonics (2004– 2008). His current interests are in high-capacity optical transmission and networking, integrated optics, and semiconductor optical amplifiers/modulators. He has (co-)authored over 150 conference and journal papers. Dr. de Waardt is member of the IEEE.