Field of the invention

The invention relates to wavelength division multiplexed (WDM) passive optical networks, and methods of operating such networks.

Background of the invention

A passive optical network (PON) is a fiber-optic point-to-multipoint network

architecture in which optical splitters enable a single optical fiber to serve multiple premises. A PON comprises an optical line terminal (OLT), for example at a service provider's central office (CO), and a number of optical network units (ONUs) near respective end users. The OLT includes a centralised light source (CLS), and the light from this source is split by at least one unpowered optical splitter to generate respective signals at the ONUs. Specifically, the light from the CLS is transmitted via a "feeder fiber" to the unpowered optical splitter, where it is split to generate light signals transmitted to the ONUs via respective "distribution fibers".

One form of PON is a wavelength division multiplexed passive optical network (WDM-PON) in which multiple wavelengths are supported in both the upstream (i.e. from ONU to OLT) and downstream (i.e. from OLT to ONU) directions, and different data is transmitted at each wavelength. In one form of WDM-PON, a separate wavelength channel is employed for downstream data transmission from the OLT to each respective ONU, and a further separate wavelength channel is employed for upstream data transmission from each respective ONU to the OLT, but other WDM schemes exist too. WDM-PON is recognized as a cost-effective and flexible solution to deliver gigabit broadband services to end-users [1-2]. Besides many inherent merits offered by WDM-PON such as virtually unlimited bandwidth, protocol transparency, excellent security, simple management, etc., the CLS configuration at the OLT together with the reuse and re-modulation of downstream signals via reflective semiconductor optical amplifiers (RSOAs) at respective optical network units (ONU), provides optical gain to the re-modulated signal and realizes colorless operation [3-4]. However, the WDM-PON architecture with CLS suffers from a critical energy consumption problem. In such a WDM-PON, even if there is no data traffic to send/receive to/from an ONU, the corresponding transmitter in the OLT still has to transmit the downstream light at the normal average optical power, so as to enable the upstream transmission, since ONU does not employ any light source and it cannot send the upstream signal without the downstream carrier. In practice, an ONU may not have upstream data in its output data queue all the time, and it may be turned off for a certain period in a day, e.g., in the early morning or at night. Hence, during the period that there is no down- and up-stream data on a certain wavelength channel simultaneously, the down- and up-stream optical power on that wavelength channel is wasted. Therefore, to reduce energy consumption, the ONU and its corresponding transceiver at the OLT should stop their signal transmission, and enter into "sleep mode" or "dozing mode", or be shut down. Recently, several energy-saving schemes for WDM-PONs were proposed [5-7]. The work in [5] used a tuneable laser to poll respective ONUs to monitor their upstream transmission status or request, while authors in [6] utilized a spectrum-sliced light-emitting-diode (LED) broadband light to carry the RF monitoring signal. In [7], the amplified spontaneous emission (ASE) spectrum from the RSOA at the ONU was modulated by a pilot tone monitoring signal, which is sent to the OLT in order to resume data transmission from the energy-saving mode. None of these schemes are cost effective and all increase the complexity of the system because of the modulation of the RF monitoring signals. Moreover, these schemes just consider one energy-saving (sleep) mode, which cannot save much energy.

Besides the energy-saving issue in a WDM-PON, the failure monitoring and protection against failures are also imperative for network operators to enhance the access network reliability. Previously proposed WDM-PONs have not considered both energy saving and failure protection simultaneously. As a WDM-PON increases the bandwidth per user to 1 Gbit/s, 10Gbit/s and beyond, any possible failure of either feeder fibers (FFs) or distribution fibers (DFs) will disrupt the services, leading to a large amount of data loss. To date, several protection schemes for WDM-PONs have been reported. In [8-9], the periodic property of arrayed waveguide gratings (AWGs) were used to enable each DF to carry data traffic for more than one ONU, while ONUs were either adjacently connected to form a group [8] or connected in sequence to form a ring [9] to offer a backup protection path. In [10], both working FF and DFs in normal mode were duplicated for protection and connected to an NxN AWG located at the remote node (RN). However, these schemes require an optical switch (OS) installed at each ONU to perform the protection switching. The distributed control scheme increases the ONU structure complexity and manufacture cost. Although some centrally controlled protection schemes were proposed in [1 1 -12], complicated OLT design and special wavelength assignment plan are required, and also N optical switches [1 1 ] or N electrical switches [12] are needed in the OLT for N ONUs. Compared with [1 1 -12], the protection scheme in [13] requires only one OS located in the OLT along with the simplified ONU structure, but it can only offer protection against FF failure.

All the existing protection schemes [8-13] described above only work under the assumption that the OLT and all the ONUs continuously transmit optical signals. However, in practice, (i) ONUs and the corresponding transceivers in the OLT may periodically enter into sleep/dozing mode whenever there is no data to be sent in order to save energy consumption, (ii) some ONUs may be turned off whenever users are offline, (iii) fiber faults may occur during the time when ONUs are in sleep (offline) mode. In the above cases, all the existing protection schemes [8-13] do not work. For example, when an ONU is in sleep (offline) mode, no optical signal is transmitted from that ONU. In such case, if one of the previous schemes is employed, since no optical signal is received from that ONU, the monitoring unit at the OLT would assume that the DF of that ONU is faulty and would trigger protection switching, resulting in a malfunction.

Summary of the invention

The present invention aims to provide a new and useful WDM-PON architecture, and in particular one which addresses at least one of the above problems. In general terms, the invention proposes a WDM-PON in which the OLT measures the power intensity of light received from each ONU, and modifies its operation based on the measured light amount.

In a first possibility, the OLT enters a resting mode (e.g. sleeping mode or dozing mode, as defined below) with respect to a given ONU when the measured light amount is below a threshold, indicating that the ONU is itself in a resting mode. The resting mode(s) of the OLT correspond to resting mode(s) of the ONUs. Which resting mode is entered may depend upon the additional factor of that there is no downstream data to be transmitted (for example within a predetermined period). This provides a simple way of performing energy saving. A simple but effective energy- saving scheme is provided by including an energy-saving control unit in the OLT and a control circuit at each ONU.

In an embodiment of the invention, the energy-saving control unit can switch the operation mode of the respective transceiver in the OLT among the active, dozing and sleep modes, by detecting the upstream light power and monitoring downstream transmission request. It not only eliminates the need for modulating any monitoring signal as in [5-7], but also requires no additional tunable supervisory transceiver as in [5] or any dedicated light source as in [6].

In a second possibility, the OLT, which is connected to a remote node (RN) via both a working feeder fiber (FF) (providing a "working path") and a protection feeder fiber (providing a "protection path"), selects which of those fibers it transmits data along, and receives data along, based on the measured light amount. A novel monitoring unit and a novel automatic protection switching control unit are provided in the OLT for centrally monitoring fiber link failures on both working and protection paths. Thus, the proposed WDM-PON architecture can deal with self-healing operation in various practical scenarios. Moreover, by simultaneously monitoring the optical power of each channel on both working and protection paths, the OLT can also accurately identify the status of every path on each channel, thus facilitating a faster failure recovery. Preferably the WDM-PON incorporates the energy-saving and self-healing operations, simultaneously. Each ONU is connected to the RN via two separate distribution fibers (DFs), which the RN respectively couples to the working FF and protection FF, and transmits any data into both of them. The OLT is configured to detect a difference in what it receives on the working path and the protection path, thereby detecting a fault. The OLT is arranged to detect and react differently to a decrease in the light amount on both paths (which may be indicative of the ONU entering a resting mode) and a decrease in the light amount on one of the paths (indicative of a fault). Thus, the WDM-PON can not only significantly reduce the energy consumption, but also deal with all possible and practical fiber failure scenarios, regardless of whether the transceivers in the ONUs are powered off or not.

Further novel features of the embodiment are listed below:

Simultaneously utilizing a transceiver unit and a monitor unit for all channels in the OLT to provide sufficient detection information to two logic decision units via an electrical power splitter: an energy-saving control unit and a protection switching control unit. The monitor unit detects the upstream signals that traverse the protection path, and the transceiver unit detects the upstream signals that traverse the working path.

A simple but effective energy-saving scheme is proposed by incorporating a logic control unit in the OLT and a control circuit in each ONU. In the energy-saving scheme, the combination of the sleep and dozing modes is utilized to further reduce energy consumption. The respective ONU is responsible for monitoring the upstream (US) data transmission state, while the OLT just detects the upstream optical power in a passive way. Hence, the scheme not only eliminates the need of modulating any monitoring signal, but also requires no additional tunable supervisory transceiver as in [5] or any dedicated light sources as in [6], which greatly reduces the network cost and complexity.

An intelligent protection switching scheme is proposed in the OLT for monitoring fiber link failures on both working and protection paths. The scheme differentiates the fiber failure case and the ONU power-off/sleep/dozing-mode case. It avoids unnecessary switching of an optical switch (OS) in the OLT. Only one 2x2 optical switch (OS) is incorporated in the OLT for the centrally- controlled protection switching. The number of the required OS is significantly reduced from N (equal to the number of ONUs) in some prior art devices, to only one. The ONU is transparent to any fiber failure and the structure of the ONU is much simplified, which also greatly reduces the amount of required network resource.

By using the protection switching control unit and two groups of the detection results on the working and protection paths, the OLT can obtain more accurate and comprehensive knowledge about each wavelength channel's state information, which facilitates a faster failure recovery.

The proposed centrally-controlled protection scheme is not restricted to a specific WDM-PON protection structure, and hence it is applicable to other WDM- PON protection structures: group protection [8], ring protection [9] and duplication protection [10]. As described above, several distributed-controlled or centrally-controlled protection schemes have been reported for survivable WDM-PONs [8-13]. However, all the previously proposed schemes assume that all the ONUs and the OLT continuously emit light, regardless of whether they have data to send or not, or if end users of ONUs are powered off or not, so they would not work if some transmitters in the ONUs or the OLT do not continuously transmit light. Also the earlier proposed distributed-controlled protection schemes require an OS installed at each ONU, and each ONU autonomously switches between the working and protection paths for bidirectional transmissions [8-10], which inevitably increases the ONU design complexity and investment cost. Although the previously proposed centrally- controlled protection schemes are improved by moving the automatic protection switching capability in the respective ONUs to the OLT located in the CO, N optical switches [1 1 ] or electrical switches [12] in the OLT are still needed for N ONUs. Moreover, complicated OLT structure design and special wavelength assignment plan are also required to perform centrally-controlled protection. In [13], the centrally- controlled protection architecture further reduces N protection OS's to only one in the OLT, which indeed simplifies OLT/ONU structure and reduces the cost. However it can only offer protection against FF failures, not the DF failures.

In the embodiment's cost-effective protection switching scheme, the above mentioned issues have been addressed. Only one 2x2 OS is incorporated in the OLT for the centrally-controlled protection switching. Such architecture not only simplifies the ONU design, but also significantly reduces the amount of required network resource. Moreover, the OLT can also keep tracking each channel's state information, thus facilitating a faster failure recovery.

In summary, preferred embodiments of the invention can not only significantly reduce the energy consumption in a WDM-PON, but also deal with all the scenarios for more practical self-healing operation. Moreover, the OLT can also keep tracking each channel's state information, thus facilitating a faster failure recovery. Using the embodiments, reliable bidirectional data transmissions in the WDM-PON can be guaranteed in a cost-effective and energy-efficient way.

Brief description of the drawings

An embodiment of the invention will now be described for the sake of example only with reference to the following drawings, in which:

Fig. 1 is a schematic diagram of an embodiment of the invention, with only one optical switch (OS) and using duplication protection topology for both distribution fibers (DFs) and feeder fiber (FFs);

Fig. 2 is composed of Figs. 2(a) which shows the state transit diagram for each ONU of the embodiment, and Fig. 2(b) which shows the state transit diagram for each channel in OLT of the embodiment; Fig. 3 is a schematic diagram of the proposed protection switching control unit in the OLT of the embodiment located at the CO;

Fig. 4 is a schematic diagram of the proposed energy-saving control unit in the ONU of the embodiment; Fig. 5 is a schematic diagram of a variant of the embodiment; and

Fig. 6 illustrates the time sequence of an energy saving operation performed by the variant of the embodiment.

Detailed description of the embodiment

A schematic diagram of an embodiment of the invention which is an energy-saving and self-healing WDM-PON architecture with N ONUs is shown in Fig.1 . An OLT located at a central office (CO) has two functional units (a "transceiver unit" and a "power monitor unit" (also referred to here as a "monitor unit")) and a control system having two logic decision units: an "energy-saving control unit" and a "protection switching control unit". These units are interconnected and co-operate to provide both effective energy-saving and intelligent protection switching. All the input and output signals of both logic decision units comply with common Transistor-Transistor Logic (TTL) level standard and are represented by logic levels "0" and "1 ". As shown in Fig.1 , the transceiver unit includes N transceivers, supporting N ONUs. In each transceiver, a transmitter (TX) generates a downstream signal and an optical circulator is used to separate down- and up-stream signals and to transfer the upstream signal to an upstream receiver (RX). Each RX also additionally acts as an optical power monitor, which detects the absence or presence of the upstream signal that traverses the working path in the normal mode. Upon detecting a drastic power loss, an electrical signal is generated and split into two copies by an electrical power splitter for two logic decision units, respectively. The wavelengths of all channels are de-/multiplexed by a 1 xN arrayed waveguide grating (AWG) in the transceiver unit of the OLT. The multiplexing port of the AWG is connected to port 1 of a 2x2 OS in the OLT. Port 3 of the 2x2 OS is connected to the monitor unit, more specifically, to the multiplexing port of another identical AWG with the same free spectral range (FSR). Each monitor (M,) in the power monitor unit detects the upstream signal traversing the protection path and likewise generates a logic electrical signal to both logic decision units via an electrical power splitter. The energy saving control unit in the OLT and the control circuit in each ONU are utilized to switch the operation mode of the transceiver in the OLT and the associated ONU in a certain wavelength channel, respectively. The protection switching control unit is used to switch the connection state (cross or bar) of the 2x2 OS in the OLT. Ports 2 and 4 of the OS are connected respectively to the multiplexing ports of two 1 xN AWGs at the remote node (RN), via two separate feeder fibers (labelled in Fig. 1 as "the working FF" and the "protection FF"). These provide respectively a "working path" and a "protection path" for up- and down-stream signals.

The RN consists of two passive 1 xN AWGs, which have the same FSR and the same operating wavelengths as the two AWGs in the OLT. After being de- multiplexed at the RN, each downstream signal is transmitted on one of the two alternate distribution fibers (DF-i and DF-i ^* , for i=1 ,...N), which connect port-i of the two AWGs to the corresponding i-th ONU (called ONU-i).

The structure of each ONU is relatively simple as shown in Fig. 1. In each ONU, a 2x2 optical coupler (OC) is used to combine the two corresponding DFs, and to split the optical power into two parts: one part is fed to a downstream receiver (RX); the other is amplified and re-modulated with upstream data via a RSOA operating in gain-saturated regime.

Energy-Saving Scheme

From the viewpoint of the RX and TX behaviour, in the proposed energy-saving scheme, three operation states are defined for the OLT and the associated ONU in a certain wavelength channel: active, sleep and dozing modes. In the active mode, both the RX and TX are turned on. The OLT and the associated ONU in a certain wavelength channel are fully responsible for the down- and up-stream traffic and consume full power P _a . The sleep mode represents the state where both the RX and TX are turned off, while just maintaining an ability to wake up on the local stimuli, for example, an upstream transmission request at the ONU, or a drastic rise of upstream power detected by the monitor at the OLT. This mode can be particularly employed when the ONU is not in use (ONU offline), or when there is no down- and up-stream traffic at the same time. Hence, the mode can achieve maximum energy reduction and only consume lowest power p _s . The dozing mode is different for each transceiver of the OLT and the associated ONU in a certain channel, respectively. The dozing mode of the ONU is referred to as ONU_dozing, and the dozing mode of the OLT as OLT_dozing. ONU_dozing only turns off the TX for a period of time (typically during periods when no upstream traffic is available), while the RX remains on. Conversely, OLT_dozing only turns off the RX, while the TX remains on. The dozing mode is very suitable for one common application scenario where the downstream traffic volume is extremely large, such as video-on-demand (VoD) and massive file download, but the upstream traffic is zero or minimal. Due to the partial shutdown of the transceiver, the power usage is reduced to a modest level P _d

( P _s < P _d < P _a ). To realize the automatic state transit in the proposed energy-saving scheme, the energy-saving control unit in the OLT and the control circuit in each ONU are designed as shown in Fig. 1. The energy-saving control unit has N identical logic modules, each of which is related to a wavelength channel. Each module has three logic input signals, which come from a downstream (DS) data queue (in the transceiver unit) ( qr, ^DS ), the upstream RX ( w, ), the power monitor ( ,), of the corresponding channel, respectively. The input logic signal gf ^s from the DS data queue indicates whether there is a DS transmission request (logic level "1 ") or not ("0"). Two logic output signals of each module are used to switch the operation mode of the RX and the TX of each wavelength channel in the OLT, respectively. As shown in Fig. 4, the control circuit in each ONU (denoted ONU,) is designed to monitor both the down- and up-stream transmission states via the detected power ( P, ^ds ) by the downstream RX, and an indicating signal ( <7," ^S ) from the upstream (US) data queue, and hence switches the operation mode of the transceiver in the ONU.

Fig. 2(a) shows the state transit diagram for one of the ONUs. The numbers in circles label four possible state transit paths. Let us assume that the ONU is initially in the state indicated as "active". As shown in Fig. 2(a), whenever the control circuit in the ONU (refer to Fig. 4) finds that there is no US signal to be sent from the US data queue for a certain period of threshold time T _th (measured using an electrical timer), a logic level "q" ^s = o " generated from the US data queue turns off the RSOA in the ONU and triggers the ONU from the active mode to the dozing mode (i.e. state transit path "T in Fig. 2(a)). Fig. 2(b) shows the state transit diagram for the OLT. Again, the numbers in circles label four possible state transit paths, and again assume that initially the OLT is in the state indicated as "active". If both the upstream RX and the monitor in the OLT detect a drastic upstream power loss, two logic signals { w, = o & p, = o) are generated to turn off the RX in the corresponding channel and to switch the OLT in the corresponding channel from the active mode to the dozing mode (i.e. state transit path "1 " in Fig. 2(b)). So the logic expression of the control signal for the upstream RX in the corresponding channel is (ιν, + ρ,), / = i,2 w (refer to Fig. 1 ), where w, and p, denote the logic detection results of the upstream optical power of channel / on the working and protection paths, respectively.

When both the ONU and the OLT in the corresponding channel are in the dozing mode, if there is also no DS data to be sent from the DS data queue, for a certain period of threshold time T _m computed by a timer, a logic signal (g, ^DS = 0 ) is generated to turn off the downstream TX of the corresponding channel in the OLT and to enter into the sleep mode (i.e. state transit path "3" in Fig. 2(b)). Thus the downstream RX in the ONU also experiences a drastic power loss due to the absence of the downstream carrier and hence generates a logic level " P,. ^DS = o " to switch the ONU into the sleep mode (i.e. state transit path "3" in Fig. 2(a)). Here the control signal for RX in the ONU is the logic OR of power detection signal P, ^os and US queue indication signal q (i.e., q" ^s + P, ^DS ) (refer to Fig. 4). It is noted that the state transit path from the dozing to the sleep mode (state transit path "3" in Figs. 2(a) and 2(b)) is an irreversible transit, and the reverse transit is prohibited, because DS data cannot be sent to a sleeping ONU, which can only be woken up only by a US data transmission request. The logic control signal R, for downstream TXi in the OLT in the corresponding channel-i is expressed as: ff,. = (q? ^s x ff,. ) ₊ ( _{Wi +} p,.) . (1 )

When the OLT in the corresponding channel is in the sleep mode, the R, = o , which removes the effect of the logic signal gf ^s on the operation mode of the downstream TX-i to avoid the above reverse transit. In other words, the signal R, will not be changed to logic "1 "; it can be changed to logic "1 " if and only if there is the upstream light received in the OLT (i.e., w, or p, or both of them are logic "1 ").

It can be also known from the Eq. (1 ) that the precondition of turning off the DS TX, in the Transceiver-i is that the US transmission has already been halted (i.e.,

" w, + p, = 0"). Therefore, in the proposed energy-saving scheme, the active mode could not be directly changed into the sleep mode and it must go through the dozing mode first.

Table 1 is a truth table for the detection states of the upstream light on both the working and protection paths in each channel. The third column of Table 1 presents the logic output of each module in the energy-saving control unit.

Table 1

When there is US data to be sent from the US data queue in a sleeping ONU or a dozing ONU, a resuming signal q? ^s (logic level "1 ") is generated by the US data queue to activate immediately the RSOA in that ONU (i.e. state transit paths "2" and "4" in Fig. 2(a)). As there is no DS signal, the RSOA is not wavelength-seeded and its broadband ASE light is sent to the OLT through both the working and protection paths. After being spectrum-sliced by the AWGs at the RN, the lights at each of the wavelengths [λ™ = λ, + mx FSR, m = ±1,±2,...) are transmitted to the OLT, due to the cyclic spectral property of the AWG. Upon the upstream optical power being detected by either the upstream RX or the monitor, or both, of the corresponding channel, the logic module in the energy-saving control unit will activate the corresponding channel to its active mode (i.e. state transit paths "2" and "4" in Fig. 2(b)), which results in turning on the upstream RX and transmitting the downstream light in the continuous wave (CW) or with data to the associated ONU. After receiving the downstream carrier again, the associated ONU resumes the normal upstream data transmission and the downstream signal receiving.

Intelligent Protection Switching Scheme

In our proposed WDM-PON architecture, an intelligent protection switching scheme is also incorporated. In the normal working mode, the 2x2 OS in the OLT is set to the bar state (i.e., 1 -2 and 3-4 connection). Thus, a downstream signal is delivered only on the working path, consisting of the working FF and respective DF-i (Red path). One part of the downstream optical power via a 2x2 OC at the corresponding ONU-i is fed to a downstream RX; the other is amplified and re-modulated with upstream data via a gain-saturated RSOA. Meanwhile, an upstream signal (which is a part of the downstream light being remodulated with upstream data via the RSOA) is generated and then is power-split by the 2x2 OC and transmitted over two different paths (working and protection) to the OLT. However, only one copy of the upstream signal transmitted in the working path will be transported to the transceiver unit via the 2x2 OS in the OLT, while the other copy on the protection path will be fed to the monitor unit via the same OS. Thus, only half of the FFs and DFs are used for data transmission, while the other half are used as backup for the normal operation mode. Therefore, the self-healing WDM-PON system can provide 1 :1 downstream protection and 1 +1 upstream protection capability, respectively. Fig. 3 schematically shows the proposed protection switching control unit in the OLT. It contains N identical logic modules, each of which is related to a wavelength channel, and a multi-input-single-output logic OR gate. For each logic module, two input signals respectively come from the upstream RX and the monitor of the corresponding channel, and its output serves as one of the N input signals of the logic OR gate. The output of the logic OR gate controls the connection state of the 2x2 OS. A single-link-failure scenario is assumed, because the chance of

simultaneous multiple-link failures is negligibly small in an access network. It is noted that the proposed protection switching scheme can also protect against

simultaneous multiple-link failures, except for the case that the two DFs for an ONU or the two FFs break down simultaneously. Under the single-link-failure assumption, considering all four possible operation scenarios, the last column of the truth table shown in Table 1 gives out the logic decision results of the protection switching control unit based on the logic inputs both on the working and protection paths in each channel.

During the normal working mode when an ONU is power-on, both the upstream RX and the monitor can detect certain light power. In the case of any working DF failure, the corresponding upstream RX in the transceiver unit will detect the loss of that upstream signal, and then a logic level "0" signal will be generated to the logic decision unit. But, in this case, a monitor in the same channel can detect light power, and a logic level "1" signal will be generated. Consequently, the output of the logic decision unit will be logic level "1" signal, which triggers the OS to the cross state (i.e., 1-4 and 3-2 connection) to setup the alternate path. Hence, all of the

bidirectional transmissions are switched from the working path (Red path) to the backup protection path (Blue path). After protection switching, the connection states of all N working paths will be monitored by N monitors in the monitor unit of the OLT. Based on the outputs of the N monitors, the monitor unit can tell if it is a DF or the FF failure in the working path; if it is a DF failure, it can also tell which DF fails. Thus a fast failure restoration can be performed. On the contrary, if an upstream RX detects the presence of light while a monitor detects no light, this indicates that the protection DF has failed (and must be repaired), but in this case no protection switching will take place. If both the upstream RX and the monitor experience a drastic power drop/loss simultaneously, then the corresponding ONU enters into sleep/dozing mode, or powers off, and again no protection switching will take place. So the proposed novel logic decision unit can distinguish between the ONU power- off/sleep/dozing-mode case and the fiber failure case, which effectively avoids unintended switching of the OS in the OLT. Thus, the logic expression of each logic module is (ΰΓ, χ-ρ, ), ; ^' = 1,2,...,W (refer to Fig. 1 ). If the above three kinds of the abnormal detection results occur in all channels, this means that the fiber failure has taken place in either the working or protection FF. A logic OR gate is used to synthetically respond to the detection states from all N logic modules. Therefore, the logic expression of the output of the whole logic decision unit is [(^ ^χ ρ, ) + (^ ^χ ₂ ) + + (^ ^χ ρ _Ν )] ·

The output signal of the logic OR gate will toggle the OS's connection state. Hence, the proposed centrally-controlled self-healing scheme in a WDM-PON can provide protection capability against the failures of both FFs and DFs.

Variants of the embodiment Many variants of the embodiment are possible within the scope of the invention. For example, Fig. 5 illustrates a WDM-PON system which is such a variant. The elements of the variant which have the same meaning as in Fig. 1 are denoted in the same way. The principle difference between the embodiment of Fig. 1 and the variant of Fig. 5 is that in Fig. 5 the energy saving control unit and the control unit of the ONU have a rather simpler structure. This structure provides just one resting state: a sleep state.

Fig. 6 illustrates the time sequence of the energy saving operation. The upper portion "1 " of the diagram illustrates a "turn-off" operation in which the WDM-PON system transitions from the active state into the sleep state with respect to a certain one of the ONUs. This happens when the ONU detects that there has been no upstream traffic for a certain threshold time T-m, and when the OLT subsequently detects that there is an absence of upstream power for that ONU. The lower portion "2" of the diagram illustrates the "turn-on" operation in which the system transitions back to the active state. This happens when the ONU detects that there is an upstream data transmission request, and the OLT subsequently detects ASE optical signals from the corresponding ONU.

Although the embodiments described above include both the energy saving feature and the protection switching feature, alternative embodiments of the invention implement just one of these two features. For example, the protection switching feature might be used in a system which does not have a resting mode at all, or which switches into or out of the resting mode by a different mechanism from the one presented here. Conversely, the energy saving feature of the embodiment can be used in a system which does not have protection switching, or which performs it by one of the methods described in the prior art.

Commercial applications of the invention

The proposed access network system can be commercialized without major modification since most of the devices and technologies in the proposed system are available in the market. Packaging and commercializing should not be an issue since the number of hardware requirements is small and the working principle is relatively straightforward.

Since all devices in the proposed system are commercially available, no special equipment is needed for manufacturing of the products. This feature also ensures that the manufacturing process is compatible with existing technologies.

Furthermore, the OLT, RN, and ONU are optimized to achieve the simplest possible design. All required components except the laser source are low-cost optical devices. Therefore, the proposed system consisting of inexpensive devices would allow cost-effective large scale production.

Telecommunication carriers and equipment manufacturers will be able to vastly enhance the energy-saving efficiency and protection capability of their access networks using our proposed systems and techniques. It is anticipated that our results will help to accelerate the practicality of the energy-saving and self-healing schemes in the WDM-PON access systems.

References [1] W. Hung, C. K. Chan, L. K. Chan, and F. Tong, "An optical network unit for WDM access networks with downstream DPSK and upstream remodulated OOK data using injection-locked FP laser," IEEE Photon. Technol. Lett., vol. 15, no. 10, pp. 1476-1478, Oct. 2003. [2] D. J. Shin, D. K. Jung, J. K; Lee, J. H. Lee, Y. H. Choi, Y. C. Bang, H. S. Shin, J. Lee, S. T. Hwang, and Y. J. Oh, "155 Mb/s transmission using ASE-injection Fabry-Perot laser diode in WDM-PON over 70 °C temperature range," Electron. Lett., vol. 39, no. 18, pp. 1331-1332, Sep. 2003.

[3] F. Payoux, P. Chanclou, M. Moignard, and R. Brenot, "Gigabit optical access using WDM PON based on spectrum slicing and reflective SOA," in Proc. Eur. Conf. Optical Commun., Sep. 2005, vol. 3, pp. 455-456.

[4] N. Nadarajah, E. Wong, K. L. Lee, and A. Nirmalathas, "Upstream access and local customer network in passive optical networks using a single wavelength seeded reflective semiconductor optical amplifier," Opt. Commun., vol. 273, no. 1 , pp. 246-251 , May 2007.

[5] T. Uchikata, and A. Tajima, "A novel power saving scheme for WDM-PON with centralized light sources", In Proc. OptoElectronics and Communications

Conference (OECC) 2009, Hong Kong, China.

[6] W. Jia, D. Shen, K.H. Tse, J. Xu, M. Li and C.K. Chan, "A novel scheme to realize a power-efficient WDM passive optical network", In Proc. OptoElectronics and Communications Conference (OECC) 2010, Japan.

[7] K.H. Tse, W. Jia, and C.K. Chan, "A cost-effective pilot-tone-based monitoring technique for power saving in RSOA-based WDM-PON", in Proc. OFC/NFOEC, Paper OThB6, Los Angeles, CA, 201 1 . [8] T. J. Chan, C. K. Chan, L. K. Chen, and F. Tong, "A self-protected

architecture for wavelength-division-multiplexed passive optical networks," IEEE Photon. Technol. Lett., vol. 15, no. 1 1 , pp. 1660-1662, Nov. 2003 [9] X. F. Sun, Z. X. Wang, C. K. Chan, and L. K. Chen, "A novel star-ring protection architecture scheme for WDM passive optical access networks,"

OFC/NFOEC, Paper JWA53, Los Angeles, CA, 2004.

[10] K. Lee, S.-G. Mun, C.-H. Lee, and S. B. Lee, "Reliable wavelength-division- multiplexed passive optical network using novel protection scheme," IEEE Photon. Technol. Lett., vol. 20, pp. 679-681 , May 2008.

[1 1] X. F. Sun, C. K. Chan, and L. K. Chen, "A survivableWDMPON architecture with centralized alternate-path protection switching for traffic restoration," IEEE Photon. Technol. Lett., vol. 18, no. 4, pp. 631-633, Feb. 15, 2006. [12] Y. Qiu, Z. X. Liu, and C. K. Chan, "A centrally controlled survivable WDM- PON based on optical suppression technique," IEEE Photon. Technol. Lett., vol. 23, no. 6, pp. 386-388, Mar. 15, 201 1.

[13] Z. X.Wang, X. F. Sun, C. L. Lin, C. K. Chan, and L. K. Chen, "A novel centrally controlled protection scheme for traffic restoration in WDM passive optical networks," IEEE Photon. Technol. Lett., vol. 17, no. 3, pp. 717-719, Mar. 2005.