WO2006119795A1

WO2006119795A1 - Method and system for optical communication comprising osnr monitoring by use of optical homodyne tomography

Info

Publication number: WO2006119795A1
Application number: PCT/EP2005/005277
Authority: WO
Inventors: Paolo Martelli; Mario Martinelli; Silvia Maria Pietralunga; Leonardo Ranzani
Original assignee: Pirelli & C. S.P.A.
Priority date: 2005-05-13
Filing date: 2005-05-13
Publication date: 2006-11-16

Abstract

An optical communication system comprises an optical-signal-to-noise-ratio monitoring apparatus including: a local oscillator; a 2x2 optical coupler adapted to mix an optical radiation from the local oscillator with a portion of an optical signal propagating through the optical communication system, said optical coupler having two output ports adapted to emit two mixed optical radiations; a balanced detector optically coupled to said two output ports, adapted to receive said two mixed optical radiations and to emit a measurement signal; a calculating device adapted to derive the optical-signal-to-noise ratio of said optical signal from said measurement signal; wherein the local oscillator is adapted to emit a continuous wave optical radiation and the optical-signal-to-noise-ratio monitoring apparatus further includes a frequency locking device coupled to the local oscillator and adapted to lock the optical frequency of the local oscillator as a function of the optical frequency of said optical signal.

Description

METHOD AND SYSTEM FOR OPTICAL COMMUNICATION COMPRISING OSNR MONITORING BY USE OF OPTICAL HOMODYNE TOMOGRAPHY

H» Η *J* "I* _"I*

Field of the invention The present invention generally relates to the field of optical communications, in particular to the field of optical communication systems including optical performance monitoring capabilities. More in particular, the invention concerns a method and a system for optical telecommunication including the functionality of optical signal to noise ratio (OSNR) monitoring by use of photon statistics reconstruction technique.

Background of the invention

A common technique to increase the transmission capacity of today optical communication systems is wavelength division multiplexing (WDM), wherein a plurality of optical carriers, each having a respective wavelength, are multiplexed together in a single optical medium, such as for example an optical fiber. The WDM channels may be closely spaced (dense WDM or DWDM) or coarsely spaced (CWDM) or a combination thereof.

Optical networking is expected to be widely used in perspective optical communication field. In the present description, the term Optical network' is referred to an optical system including a plurality of point-to-point or point-to-multipoint (e.g., ring) optical systems optically interconnected through nodes. In all-optical transparent networks few or no conversion of the optical signal into electrical signal, and then again in optical signal, occur along the whole path from a departure location to a destination location. This is accomplished by placing at the nodes of the optical networks electro- optical devices which are apt to process the optical signal in the optical domain, without the need for electrical conversion. Examples of such devices are OADM, branching unit, optical router, optical switches and the like. In addition, optical systems, and at a greater extent optical networks, make use of optical amplifiers in order to compensate the power losses due to fiber attenuation or to insertion losses of the optical devices along the path, avoiding the use of any conversion of the optical signal into the electrical domain even for long traveling distances and/or many optical devices along the path.

In optical systems, and at a greater extent in optical networks, a problem exists of monitoring the optical performance of the optical signal. For the management of an optical system, monitoring the optical signal to noise ratio (OSNR) is particularly important. This is especially true in WDM systems, and at a greater extent in WDM networks, wherein the various channels can travel through different optical paths and 5 have therefore different OSNRs at a given location (e.g. at the receiver). Developing of new techniques to monitor the OSNR when optical fibers and amplifiers are present is therefore of special interest.

The optical noise spectrally superimposed to the optical signal, which may degrade the transmission quality, may have various sources (such as amplified 0 spontaneous emission or ASE, relative intensity noise or RTN of the optical source, multiple path interference or MPI, modulator generated noise, etc.) and characteristics. In optically amplified system, typically the ASE represents the principal source of the optical noise.

For a given physical optical signal, the measured value of OSNR depends on the 5 definition used for the OSNR, which in turn typically is operatively based on the technique used for the measurement. More details on this will be given in the description further below.

Conventional techniques make use of a spectrum analyzer (either optical or electrical) to measure OSNR, but they are not adapt to measure directly the noise at the O signal optical frequency (in-band noise) and they become inaccurate in case the signal has low power levels. Moreover optical spectrum analyzers have a low spectral resolution, while electrical ones have large electronic noise. Finally spectrum analyzers are in general bulky and expensive. These characteristics make these techniques not suitable for the in-line monitoring of the performance of an optical link. 5 Other OSNR monitoring techniques have been proposed which do not make use of optical or electrical spectrum analyzer.

For example, narrow-band notch filter may be utilized to remove a component of the signal so that the remaining signal can be measured. Two detectors may be utilized with the power in the channel being measured by a low-gain detector and the power in O the noise being measured by a high-gain detector. This solution has the disadvantage that it cannot measure the noise power spectral density at the signal frequency, but only next to it (out-of-band noise). This can cause errors especially in WDM transmission, because the noise next to one WDM channel is influenced by adjacent channels too. Some methods have been proposed to measure the noise spectral density at the signal frequency based on the different characteristics of the signal and the noise.

For example, a method for monitoring the OSNR using a polarization-nulling method may comprise the steps of: linearly polarizing an arbitrarily polarized optical signal including an unpolarized ASE noise; separating the optical signal and the ASE noise from the linearly polarized optical signal including the unpolarized ASE noise to measure a power of the optical signal and a power of the ASE noise included in a bandwidth of an optical signal; and obtaining the OSNR using the measured optical signal power and ASE noise power. Alternatively the orthogonal delayed-homodyne technique estimates the OSNR by analyzing the receiver noise characteristics after eliminating the signal components at a specific frequency within the modulation bandwidth, hi this scheme the signal is split into two orthogonal polarized components. hi a further alternative, the optical signal together with the optical noise transmitted therewith is fed to an optical filter. The optical output signal from the above is converted into a corresponding electrical signal in a detector device and either the mid-frequency of the optical filter or the detector device is periodically modulated with a modulation signal. The received total light power is determined from a direct current component of the electrical signal and the optical signal power is determined from a time-dependent modulation component. The carrier-to-noise ratio is determined from the above parameters.

Among the above techniques, those which require narrow optical filtering in order to select the optical channel under testing have the disadvantage of increased complexity and costs of OSNR monitoring, in that typically a narrow-band optical filter requires wavelength stabilization. A narrow-band optical filter will be referred to as an optical filter having a bandwidth of the order of the WDM wavelength spacing, i.e. typically below about 1 nm or less. For example for a 10 Gbit/s WDM signal having 0.2 nm wavelength spacing, the optical filter needs a bandwidth of less than about 0.2 nm.

Moreover all the above techniques show the problem of the electrical noise of photoreceiver(s), which, in order to monitor the OSNR of a signal, results in the need of tapping an amount of signal power which would cause a noticeable subtraction of power from the optical signal propagating along the line, thus disturbing the transmission.

The article 'Linear Optical Sampling' by C. Dorrer et al., IEEE Photonics Technology Letters, Vol. 15, No. 12, December 2003, pages 1746-8 describes the measurement of waveforms and eye diagrams at high bit rates by optical sampling using coherent detection. Open eye diagrams could be measured at data average powers down to -20 dBm. m the art, the technique of optical homodyne tomography (OHT) has been applied in the optical communication field to characterize or test optical devices. For example, the article 'Photon Statistics of Single-Mode Zeros and Ones from an Erbium Doped Fiber Amplifier Measured by Means of Homodyne Tomography' by P. Voss et al., IEEE Photonics Technology Letters, Vol. 12, No. 10, October 2000, pages 1340-2, describes an experimental validation of the theoretical description provided by a quantum- mechanical model of ASE noise statistics emerging from an erbium-doped fiber amplifier (EDFA). The authors apply OHT to show that the zeros (no input to the amplifier) in a single mode of an EDFA are Bose-Einstein distributed and the ones (coherent-state input to the amplifier) are Laguerre distributed. To obtain the coherent- state input to the EDFA, a 95/5 fiber coupler was used after the local oscillator (LO) laser. The 95% output was used for the LO and the 5% output as the input light to the EDFA. By disconnecting and connecting input light to the EDFA, ASE (zeros) and amplified coherent-state light (ones) respectively, were obtained. The resulting noise figure of the amplifier compares well to that measured by an optical spectrum analyzer. The Applicant has found that it is highly desirable to provide a system and method for optical transmission with a functionality of on-line OSNR monitoring which do not degrade the optical transmission quality and which allow the monitoring of OSNR in principle in any arbitrary position along the optical line. In order to achieve this goal, it is advantageous that the on-line OSNR monitoring be characterized by a very high sensitivity, i.e., it allows the measurement of the OSNR of very weak signal power tapped from the optical line of the transmission system. This feature provides great flexibility in the design of an optical transmission system in that the functionality of on-line OSNR monitoring may be designed virtually at any arbitrary point along the line, thus removing the constraint of selecting those points along the line wherein the optical signal propagating therethrough is above a certain threshold, hi this way, a system designer may place the OSNR monitoring functionality also where the signal is low, such as at the end of a fiber transmission span, before optical amplification or detection. Typically, the average signal power per each wavelength along the optical line reaches a maximum of about 0 dBm to 10 dBm (typically downstream an optical amplifier) and a minimum of about -40 dBm to -20 dBm (typically upstream an optical amplifier such as a pre-amplifϊer). Assuming an optical tap which extract about 10% of the optical signal (90/10), this turns in a tapped optical signal power as low as about - 50 dBm. Using a 99/1 optical tap, the tapped signal is 20 dBm lower than the propagating signal, down to -60 dBm.

Accordingly, the Applicant has found a new need of on-line OSNR monitoring functionality with sensitivity below -30 dBm, preferably below -40 dBm, at least for OSNR values which are of the order of magnitude typical in the optical transmission field.

For the purpose of the present description, the term sensitivity of an in-line optical performance monitoring method and apparatus will refer to the minimum optical signal power tapped from the optical line which allows an accurate OSNR measurement. This value may depend on the actual OSNR value.

Summary of the invention

The invention relates to a method and a system for optical transmission furnished of in-line OSNR monitoring which can solve the problems stated above. In particular this method and system rely on OSNR monitoring which is capable of measuring the noise power spectral density at the signal frequency, is sensitive to very low power levels, and avoids the use of a narrow optical filter to select the optical channel under testing. The OSNR monitoring may be placed at any arbitrary point along the optical link. The solution of the present invention is simple, feasible and low cost. The low power level sensitivity allows constant in-line monitoring of the transmitted signal by tapping a small portion of it without stopping or disturbing the transmission of the data at any location along the optical line. Moreover the OSNR monitoring apparatus of the present invention can be realized completely in fiber to avoid alignment problems and related high costs.

The Applicant believes that the OSNR monitoring technique according to the present invention achieves some or all of the above advantages being based on the reconstruction of the photon statistics of the optical signal and on the calculation of the OSNR as the ratio between the squared photon mean and the photon variance. In one embodiment, this method reconstructs the photon statistics from the statistical distribution of the signal quadrature, i.e. the real part of the complex amplitude, when the difference between the signal phase and the local oscillator phase has a uniformly distributed random variation with a characteristic time longer than the detection time (reciprocal of the electrical bandwidth) of the receiver. The quadrature is measured by balanced detection. In this detection scheme the signal is mixed in a 50/50 optical coupler with a local oscillator, which is a laser having advantageously a much higher power than the signal. The coupler outputs may be directed to two (e.g. p-i-n) photodetectors and the generated photocurrents may be amplified and subtracted. The resulting electrical signal is typically proportional to the optical signal quadrature relative to the local oscillator phase and multiplied by the local oscillator amplitude. Advantageously this electrical signal depends only on the signal and noise components within the receiver bandwidth around the signal frequency. In order to reconstruct the photon statistics, the difference between the signal phase and the local oscillator phase are preferably scrambled uniformly in time. According to the present invention this may be obtained in that the signal and the local oscillator come from different sources and hence their intrinsic phase fluctuations are uncorrelated. This may result in a relative phase difference varying uniformly in time. The characteristic time of the phase difference fluctuations is preferably much higher than the receiver detection time. This can be achieved when both signal and local oscillator linewidths are much lower than the receiver bandwidth. In this case, the phase difference between signal and local oscillator remains substantially constant during the detection time interval. For the purpose of the present invention, the signal linewidth is the intrinsic linewidth of the optical carrier emitted by the transmitter's laser source before data modulation. The local oscillator frequency is controlled by a frequency locking mechanism in order to compensate the relative frequency fluctuations between the signal and the local oscillator and keep the electrical signal into the receiver band. Once this has been done, the local oscillator selects one signal channel without the need of any narrow optical filtering of the WDM channel.

The new approach of the present invention gives the possibility to characterize a signal taken from a real optical communication channel, which is not a portion of the local oscillator.

The frequency locking can be achieved according to the various techniques known in the field. In a first exemplary realization of the present invention the locking scheme comprises a frequency to voltage converter. According to this scheme the beat signal between the lasers to be locked is measured. A frequency to voltage converter generates a voltage proportional to the frequency of this beat signal. The voltage is compared with a reference to generate an error signal, which is used to control the frequency of one of the two lasers.

In a second exemplary realization of the present invention a different locking mechanism is used which is similar to the previous one, but a large dynamic phase comparator is used instead of the frequency to voltage converter. Thanks to the large dynamic of the phase comparator, the loop can keep the laser frequencies locked even if the relative phase between the lasers is not controlled completely and still fluctuates.

In an aspect the present invention relates to an apparatus to monitor the OSNR in a fiber optic communication link. This apparatus receives a small portion of the optical power tapped from a point of the fiber link without disturbing appreciably the transmission and reconstructs the photon statistics. The OSNR is then calculated as the ratio between the squared photon mean and the photon variance. Among the advantages the Applicant has found the great sensitivity to very low power levels and the ability of selecting a single optical channel without using a narrow optical filter. These advantages are related to a continuous wave local oscillator which both amplifies the signal and selects one optical channel centered at the local oscillator frequency. A continuous wave local oscillator also shows the advantage of having a lower cost when compared to a pulsed local oscillator. The apparatus of the present invention allows the measurement of the noise power spectral density at the signal frequency, giving more accurate results. The Applicant believes to have realized for the first time a photon statistics reconstruction apparatus in all-fiber configuration, as compared to free-space experimental setups. As a consequence, the apparatus for OSNR monitoring is more robust and more portable and suitable to make measures on the field. The setup according to the present invention may advantageously be used to characterize the OSNR of intensity modulated signals independently by the modulation format and rate, when the modulation rate (typically in the Gbit/s order of magnitude) is much higher than the receiver bandwidth (in the order of MHz). The intensity modulation of the signal generates high frequency components which are filtered by the apparatus of the present invention while the optical carrier is characterized. According to an aspect of the present invention, an optical communication system as set forth in the appended claim 1 is provided. The optical communication system comprises an optical-signal-to-noise-ratio monitoring apparatus including: a local oscillator; a 2x2 optical coupler adapted to mix an optical radiation from the local oscillator with a portion of an optical signal propagating through the optical communication system, said optical coupler having two output ports adapted to emit two mixed optical radiations; a balanced detector optically coupled to said two output ports, adapted to receive said two mixed optical radiations and to emit a measurement signal; a calculating device adapted to derive the optical-signal-to-noise-ratio of said optical signal from said measurement signal; wherein the local oscillator is adapted to emit a continuous wave optical radiation and the optical-signal-to-noise-ratio monitoring apparatus further includes a frequency locking device coupled to the local oscillator and adapted to lock the optical frequency of the local oscillator as a function of the optical frequency of said optical signal. Preferably, the calculating device is adapted to derive said optical-signal-to-noise ratio via an algorithm suitable to derive, from said measurement signal, statistical information related to the photon number probability distribution of said portion of optical signal. In one embodiment, said statistical information comprises the photon number probability distribution of said portion of optical signal. hi another embodiment, said statistical information comprises statistical moments of the photon number probability distribution of said portion of optical signal, wherein, preferably, said statistical moments comprise the mean and the variance of the photon number probability distribution of said portion of optical signal. Advantageously, the measurement signal contains information on the optical signal quadrature in-phase with the optical radiation from the local oscillator. Preferably, the calculating device is adapted to reconstruct the phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator, hi an embodiment, the calculating device is adapted to derive said optical-signal-to- noise ratio via an algorithm suitable to derive the photon number probability distribution of said portion of optical signal via a pattern functions weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator. In another embodiment, the calculating device is adapted to derive said optical-signal- to-noise ratio via an algorithm suitable to derive the statistical moments of the photon number probability distribution of said portion of optical signal via a Hermite polynomial weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phasς with the optical radiation from the local oscillator. The optical communication system advantageously further comprises an optical tapping device configured to extract said portion of the optical signal propagating through the optical communication system and to input said portion to the optical coupler.

According to another aspect of the present invention, a method for optical communication as set forth in claim 11 is provided. The method comprises monitoring the optical-signal-to-noise-ratio of an optical signal according to the following steps:

- tapping a portion of said optical signal;

- mixing said portion of optical signal with an optical radiation;

- splitting said mixed optical radiation into two output radiations; - receiving said two output radiations and emitting two corresponding electrical signals;

- emitting a measurement signal as a function of the difference of said two electrical signals;

- deriving the optical-signal-to-noise-ratio of said optical signal from said measurement signal; wherein said optical radiation is a continuous wave and the method further comprises the step of locking the optical frequency of said optical radiation as a function of the optical frequency of the optical signal.

Preferably, said optical-signal-to-noise-ratio is derived using an algorithm suitable to derive, from said measurement signal, statistical information related to the photon number probability distribution of said portion of optical signal.

In one embodiment, said statistical information comprises the photon number probability distribution of said portion of optical signal. In another embodiment, said statistical information comprises statistical moments of the photon number probability distribution of said portion of optical signal, wherein, preferably, said statistical moments comprise the mean and the variance of the photon number probability distribution of said portion of optical signal.

Advantageously, in the methods above the measurement signal contains information on the quadrature of said portion of optical signal in-phase with said optical radiation. Preferably, deriving the optical-signal-to-noise-ratio comprises reconstructing the phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation. According to a still further aspect of the present invention, an apparatus for monitoring the optical-signal-to-noise-ratio of an optical signal as set forth in claim 18 is provided. The apparatus includes: a local oscillator; a 2x2 optical coupler having a first input port optically coupled to the local oscillator and a second input port adapted to receive the optical signal, the optical coupler being adapted to mix an optical radiation from the local oscillator with the optical signal and further having two output ports adapted to emit the mixed optical radiation in two portions; a balanced detector optically coupled to said two output ports, adapted to receive said two portions of the mixed optical radiation and to emit a measurement signal; a calculating device adapted to derive the optical-signal-to-noise-ratio of said optical signal from said measurement signal; wherein the local oscillator is adapted to emit a continuous wave optical radiation and the apparatus further includes a frequency locking device coupled to the local oscillator and adapted to lock the optical frequency of the local oscillator as a function of the optical frequency of the optical signal. Preferably, the calculating device is adapted to derive said optical-signal-to-noise-ratio via an algorithm suitable to derive, from said measurement signal, statistical information related to the photon number probability distribution of said portion of optical signal.

In one embodiment of the apparatus, said statistical information comprises the photon number probability distribution of said portion of optical signal. In another embodiment of the apparatus, said statistical information comprises statistical moments of the photon number probability distribution of said portion of optical signal, wherein, preferably, said statistical moments comprise the mean and the variance of the photon number probability distribution of said portion of optical signal. Preferably, the measurement signal contains information on the optical signal quadrature in-phase with the optical radiation from the local oscillator. More preferably, the calculating device is adapted to reconstruct the phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator. In one configuration, the calculating device is adapted to derive said optical-signal-to- noise ratio via an algorithm suitable to derive the photon number probability distribution of said portion of optical signal via a pattern functions weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

In another configuration, the calculating device is adapted to derive said optical-signal- to-noise ratio via an algorithm suitable to derive the statistical moments of the photon number probability distribution of said portion of optical signal via a Hermite polynomial weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

Brief description of the drawings

The features and advantages of the present invention will be made clear by the following detailed description of an embodiment thereof, provided merely by way of non-limitative example, description that will be conducted making reference to the annexed drawings, wherein:

Figure 1 schematically shows in terms of functional blocks an exemplary optical communication system architecture according to the present invention;

Figure 2 is a schematic diagram showing in terms of functional blocks an exemplary configuration of an optical performance apparatus according to the present invention; and

Figure 3 is a schematic diagram showing in terms of functional blocks an exemplary alternative configuration of the optical performance apparatus according to the present invention.

Detailed description of the preferred embodiment(s) of the invention

Figure 1 shows an optical communication system architecture according to a possible embodiment of the present invention.

The optical communication system 100 comprises at least a transmitter 110, a receiver 120 and an optical line 130 which optically connects the transmitter and the receiver. The transmitter 110 is an opto-electronic device apt to emit an optical signal carrying information. It typically comprises at least an optical source (e.g., a laser) apt to emit an optical radiation and at least a modulator apt to encode information onto the optical radiation. Preferably, the transmitter 110 is a WDM transmitter (e.g., either DWDM or CWDM) and the optical signal comprises a plurality of optical wavelengths each carrying modulation-encoded information. The receiver 120 is a corresponding opto-electronic device apt to receive the optical signal emitted by the transmitter and to decode the carried information. The optical line 130 may be formed by a plurality of sections of optical transmission media, such as for example optical fiber sections, preferably cabled. Between two adjacent sections of optical fiber, an optical or optoelectronic device is typically placed, such as for example a fiber splice or a connector, a jumper, a planar lightguide circuit, a variable optical attenuator or the like. For adding flexibility to the system 100 and improving system functionality, one or a plurality of optical, electronic or opto-electronic devices may be placed along the line 130. In figure 1 a plurality of optical amplifiers 140 are exemplarily shown, which may be line-amplifier, optical booster or pre-amplifier. Also, a couple of optical add and drop multiplexers (OADMs) 150 are exemplarily shown along the optical line 130, which are apt to add and/or drop one or more optical wavelength(s) propagating through the optical line. Preferably, the OADMs are configurable, even dynamically, or, more preferably, tunable. The dropped or added wavelength(s) may be received or transmitted, respectively, by further receiver(s) 152 or transmitter(s) 154 which may be co-located with the OADM node or at a distance thereof. An optical system 100 having add-drop nodes, as shown in Figure 1, is commonly referred to as an optical network and it is characterized by having a plurality of possible optical paths for the optical signals propagating through it. As exemplarily shown in Figure 1, a number of six optical paths are in principle possible, which corresponds to all possible choices of the transmitter-receiver pair in Figure 1 (excluding the pairs belonging to the same node).

According to the present invention, the optical system 100 comprises at least one optical performance monitoring device 200 optically coupled to the optical line 130 at any arbitrary location by an optical tapping device 210 and apt to monitor the optical performance of the optical signal propagating at said location. The optical tapping device 210 is apt to extract from the optical line 130 a portion of the optical signal propagating therethrough and to input it into the monitoring device 200. Preferably, the optical tapping device 210 is apt to extract a small portion of the optical signal, for example less than 10% in optical power, preferably less than 5%, more preferably less than 1%. The optical tapping device 210 may advantageously be an optical fiber coupler of the kind, e.g., 90/10, 95/5 or 99/1. Alternatively, the optical tapping device 210 may be a planar lightguide circuit optical coupler or a micro-optics device containing a polarization beam splitter. Preferably, the optical tapping device 210 is a broadband coupler, spanning at least 30 nm (from 1530 to 1560), preferably at least about 100 nm (1520-1620), more preferably at least 300 nm (from about 1300 nm to about 1600 nm). The optical system or network 100 may advantageously comprise a plurality of monitoring devices 200. In Figure 1 a couple of further monitoring devices 200' and 200" are exemplarily shown, together with their respective optical tapping devices 210' and 210".

According to the present invention, the monitoring device 200 is based on a technique which reconstructs the photon statistics of the optical signal.

Figure 2 shows a schematic diagram of a monitoring apparatus 200 in accordance with an embodiment of the present invention. The monitoring apparatus 200 comprises a photon statistics reconstructing device 300 and an optical frequency locking device 400. Devices 300 and 400 are functional devices: they may or may not correspond to physical devices and they may or may not be physically distinct, hi particular it is noted that functional block 8 (described below as part of device 300) may be viewed as functionally belonging also to device 400. hi one embodiment (not shown), two distinct functional blocks 8 may be comprised in the apparatus 200, one belonging to the device 300 and the other to the device 400.

The photon statistics reconstructing device 300 comprises a local optical oscillator 1 (e.g. Photonetics Tunics-1550), a 2x2 optical coupler 4 having a first input port optically connected to the local oscillator, a second input port and two output ports, and a balanced detector 5 optically connected to the two output ports of the optical coupler 4 and having an output port. The local oscillator 1 is adapted to emit a continuous wave optical radiation, which means having a substantially constant optical power during the single datum acquisition measurement time. Typically, it is substantially constant for a length of time grater than about 1 μs.

An input port 50 of the monitoring apparatus 200 is optically connected to the second input port of the optical coupler 4 and it is adapted to receive a portion of the optical signal propagating through the optical system 100 via the optical tapping device 210 of Figure 1.

The 2x2 optical coupler 4 is any system having two input ports and two output ports, wherein any output port is coupled to both the input ports and wherein the coupler 4 is configured to give to that part of optical radiation inputting an input port and outputting the respective cross output port a π/2 phase shift with respect to that part of optical radiation inputting the same input port and outputting the respective bar output port. The terms 'cross' and 'bar' are used in the meaning common in the art. Preferably, the 2x2 optical coupler 4 is balanced, i.e. it is configured to split an optical radiation inputting into one of the two input ports into two optical radiations having equal optical power and exiting the two output ports respectively. The optical coupler 4 may advantageously be a planar lightguide circuit 50-50 optical coupler or a fused fiber 3dB optical coupler or a beam splitter. Preferably, it is a broadband (at least 30 nm or 100 nm) 50-50 optical coupler. The optical coupler 4 may show a wavelength dependency, e.g. the splitting ratio of the two output radiations may deviate, at a given wavelength, from the balanced working point described above. In this sub-optimal case, the apparatus 200, while still working according to the principle of the present invention, shows degradation in the sensitivity.

The balanced detector 5 comprises a pair of photodiodes 5' and 5", such as for example p-i-n photodiodes, and a differential amplifier 6 connected to both the photodiodes. The photodiodes 5' and 5" typically comprise a respective transimpedance amplifier for converting the generated photocurrent in voltage. The photodiodes 5' and 5" exhibit the same behavior. For example, they have the same responsivity and the two transimpedance amplifiers have the same transimpedance gain. Preferably they are identical photodiodes. They may exemplarily be germanium photodetectors with 300 MHz bandwidth and 0.6 pW/sqrt(Hz) net equivalent noise power (NEP) at the input.

A computing device 10, such as a computer, is operatively connected to the output port of the balanced detector 5 and it is adapted to numerically process the measurement data received from the balanced detector.

A plurality of electrical devices (7, 8, 9, 30) are advantageously operatively associated to the balanced detector 5 and are adapted to electrically process the measurement data which are emitted from the balanced detector 5 before the numerical processing by the computer 10, so as to reduce the electrical noise generated by the balanced detector 5. In Figure 2 it is exemplarily shown a band pass filter 7 connected to the output port of the balanced detector 5. A radio frequency (RF) signal generator 8, which is preferably adapted to generate a RF sinusoidal wave having advantageously a frequency which ranges from 2 MHz to 100 MHz (exemplarily 12.4 MHz), is advantageously connected via a multiplier 30 to the band pass filter 7. The band-pass filter 7 is preferably centered at substantially the same frequency of the RF signal generator 8 and has a passband whose width ranges from about 1 MHz to about 50 MHz. Exemplarily it is centered at about 12.4 MHz and has a 4 MHz bandwidth. A low pass filter 9 is advantageously interposed between the multiplier 30 and the computing device 10. A polarization controller 12 is preferably optically interposed between the local oscillator 1 and the optical coupler 4 and it is adapted to control the polarization of the optical radiation emitted by the local oscillator. A further polarization controller 13 is preferably optically interposed between the input port 50 of the monitoring apparatus 200 and the optical coupler 4 and it is adapted to control the polarization of the portion of optical signal under monitoring.

Preferably an optical filter (OF) 15 is optically interposed between the input port 50 of the monitoring apparatus 200 and the optical coupler 4 and it is adapted to advantageously filter out the portion of optical radiation (both signal and noise) spectrally far from the optical signal under measurement. The optical filter 15 is adapted to advantageously reduce the total power incident on the receiver 5 so as to avoid its saturation. Accordingly, the optical filter 15 preferably has a bandwidth greater than the WDM wavelength spacing, i.e. typically grater than about 0.5 nm or about 1 run, and it advantageously does not need stabilization.

Preferably a variable optical attenuator (VOA) 14 is optically interposed between the input port 50 of the monitoring apparatus 200 and the optical coupler 4 and it is adapted to advantageously control the signal power at the optical coupler 4.

Advantageously an optical isolator 11 is optically interposed between the local oscillator 1 and the optical coupler 4 and a further optical isolator 11' is optically interposed between the input port 50 of the monitoring apparatus 200 and the optical coupler 4.

The Applicant has found that, in an all-fiber configuration of the photon statistics reconstructing device 300, the optical backreflection from the components such as couplers, connectors and the like, generates one or more replica of the optical radiations, thus degrading unacceptably the measure. The Applicant has found that inserting a couple of isolators 11 and 11' as shown in Fig 2 solves this problem. The Applicant has recognized that the insertion of the two isolators 11 and 11' in an all-fiber photon statistics reconstructing device 300 is advantageous also when the all-fiber device 300 is used, as a self standing apparatus without the frequency locking device 400, for off-line characterization of optical devices.

The frequency locking device 400 is optically connected both to the local oscillator 1 and to the input port 50 of the performance apparatus 200 so as to receive a portion of the optical radiation, respectively, emitted by the local oscillator 1 and in input to the input port 50. Exemplarily, Figure 2 shows an optical coupler 2 optically connected to the local oscillator 1 and adapt to extract a preferably small portion of the optical radiation emitted by the latter. For example, the extracted portion is less than or equal to about 10%, more preferably less than or equal to about 5%. Exemplarily, Figure 2 shows a further optical coupler 3 optically connected to the input port 50 and configured to extract a preferably large portion of the optical signal in input to the apparatus 200 through the input port 50. For example, the extracted portion is greater than or equal to about 80%, more preferably greater than or equal to about 90%.

The frequency locking device 400 is operatively connected to the local oscillator 1 so as to drive the optical frequency of the latter in response to the received two portions of optical radiation. The frequency locking device 400 may be any device which is adapted to lock the optical frequency of the local oscillator 1 as a function of the optical frequency of the optical signal received by the apparatus 200 from the optical system 100 of Figure 1.

Figure 2 exemplarily shows one of the possible realizations of the frequency locking device 400. Accordingly, an optical coupler 16 has two input ports optically connected to the couplers 2 and 3, respectively, so as to be adapted to receive and mix together the extracted portion of the optical radiation emitted by the local oscillator 1 and the extracted portion of the optical signal in input to the apparatus 200. Preferably, an optical isolator 17 and a polarization controller 18 are optically interposed between the coupler 2 and the coupler 16. In addition or alternatively, the polarization controller

18 may be optically interposed between the coupler 3 and the coupler 16. Preferably a further isolator 17' is optically interposed between the coupler 3 and the coupler 16. Optical coupler 16 has an output optical port. Optical coupler 16 may exemplarily be a 50-50 optical coupler or an Y-branch coupler.

A photodetector 19 is optically connected to an output port of the coupler 16. A voltage comparator 20 is apt to receive the output (preferably amplified) of the photodetector 19 and to compare it with a reference voltage (e.g. the mean value of the signal in output from photodetector 19). A frequency divider 21, a frequency-to-voltage converter (FVC) 22, an electrical comparator 23, an electrical integrator 24 and a variable electrical attenuator (VEA) 28 are connected in series downstream the voltage comparator 20.

A radio frequency (RF) signal generator 8, a frequency divider 25 and a further FVC 26 are series-connected to a second input port of the comparator 23. The same RF signal generator 8 of the device 300 is advantageously used.

In operation, the optical performance monitoring apparatus 200 receives at its input port 50 a fractional portion S_IN of the optical signal propagating through the optical system 100 of figure 1. The optical radiation emitted by the local oscillator 1 and the input optical signal

S_IN are made to interfere in the 2x2 optical coupler 4. The optical radiation of the local oscillator is preferably much stronger than the optical signal at the input of the coupler 4, for example it is at least about 10 times or at least about 100 times stronger than the optical signal. The mixed optical radiation and optical signal is divided into the two output ports of the optical coupler 4, and inputted to the two photodiodes 5 and 5". Their currents are then converted to voltage signals and subtracted in the differential amplifier 6. After that the signal is preferably made to pass through the electrical bandpass filter 7, multiplied by a sinusoidal wave generated by the RF signal generator 8 and low-pass filtered by the filter 9. The band-pass filter 7 removes the non-white electronic noise such as for example the so called "Vf electronic noise" which has a power spectral density inversely proportional to the spectrum frequency / and is thus present at low frequencies. The low-pass filter 9 removes the electrical noise out of the band of interest.

The resulting electrical measurement signal (which is typically a voltage V) is then passed to the calculator 10 in order to be numerically processed.

First, the resulting measured signal, given by V, is normalized according to the formula:

V wherein ΔV_vacuum is the standard deviation of the resulting measurement signal when only the radiation from the local oscillator 1 (and no optical signal) is fed into the coupler 4 and V is the value measured when the optical signal is present.

The value of q so obtained represents the optical signal quadrature in-phase with respect to the optical radiation from the local oscillator normalized to the vacuum and may be represented by: q=qscos(φ_Oi)+Pssin(φoi)=^zA_scos(φs-φoi) (1) wherein qs and ps are the so called signal quadratures of the signal field given by:

Es=A^ =q_s+ips and φoi is the phase of the local oscillator field given by:

Eo_L=Ao_Le^iφOL Any measured value of q refers to a value of φs -Φ_OL-

Preferably, the local oscillator and the optical signal have the respective litiewidth sufficiently lower than the receiver bandwidth in order to guarantee that φs - Φ_OL remains substantially constant along the measurement time interval, which is about the inverse of the electrical bandwidth of the receiver (where the electrical bandwidth of the receiver is the overall bandwidth of the whole receiving set-up, including balanced detector 5, filters 7 and 9, mixer 30 and the acquisition board of the electronic calculator 10). For example the electrical bandwidth of the receiver may be about 2 MHz and the linewidth of the local oscillator and of the optical signal less than about 1 MHz, e.g. about 100 kHz. On the other hand, in case both the two linewidths are very narrow, a very long time should be needed in order to have a phase scrambling sufficient to reconstruct the photon statistics. This sets a lower limit for the narrower of the two linewidths. Preferably, the optical power of the local oscillator 1 (related to Aoi) is substantially constant along each sample q measurement time interval. This is also refereed to as a continuous wave local oscillator. Typically, it is substantially constant for a period grater than about 1 μs.

The computer 10 progressively acquires the data of q during the scrambling of Φ_S-Φ_OL and generate the phase-averaged statistical distribution of the quadrature values pr(q). Statistical information related to the photon number probability distribution of the portion of optical signal is then obtained by using any algorithm known in the art which is suitable to derive the photon number distribution p(n) or its statistical r-order moments starting from the statistical distribution of the phase-averaged quadrature values pr(q). Exemplarily, the photon distribution p(n) may be derived by the statistical average of the so called amplitude pattern functions f_m(q) '■

+∞ Pin) =< f_m (q) >\_g = \f_m (q)pr(q)dq (2)

wherein «=0, 1, 2, etc. For the purpose of the present invention, eq. (2) will be referred to as 'pattern functions weighted average' of the phase-averaged statistical distribution of the optical signal quadratures in-phase with respect to the optical radiation from the local oscillator /?rfø).

The amplitude pattern functions are known functions given by:

dq where ψ_n(Φ are the normalizable solutions of the Schroedinger wave equation of the harmonic oscillator:

and the φ_n(q) are the unnormalizable solutions of the same equation. Once the photon statistics p(n) is computed according to eq. (2), the photon mean <n> and the variance <Δn²> are obtained from the reconstructed photon statistics:

< n > = ∑np(n)

(3). < An² > = ^ (n- < n >)² p(n)

B=O

Alternatively, in order to calculate any factorial moment of order r of the photon distribution, («^(r) ) , without deriving p(n), it is possible to use the following integral:

(4)

where H_2r(q) is the Hermite polynomial of order 2r evaluated in correspondence of the phase-averaged quadrature q. For the purpose of the present invention, eq. (4) will be referred to as the Ηermite polynomial weighted average' of the phase-averaged statistical distribution of the optical signal quadratures in-phase with the optical radiation from the local oscillator pr(q).

Once the photon mean <n> and the variance <Δn > are obtained from eq. (3) or from the first (r=l) and the second (r=2) moment from eq. (4), then the OSNR of the optical signal may be evaluated by the formula:

OSNR = ^{< n >}, (5).

< An >

Eq (5) represents the definition of OSNR for the purpose of the present invention. When using an optical spectrum analyzer in case of optically amplified signals, in which the noise is principally related to the ASE, an operative quantity

is measured, wherein P_si_g and PASE are the optical power of, respectively, the signal and of the amplified spontaneous emission (ASE) in a reference optical bandwidth B_opt (typically 0.1 nm) and wherein the P_ASE is taken near the signal peak (it is assumed the relevant source of noise is ASE). hi particular, in the case of dominant contribution to the variance at the denominator of eq. (5) given by the signal- ASE beating, the OSNR* measured by an OSA is directly proportional to the OSNR according to eq. (5).

Now, the sensitivity of the photon statistics reconstruction device 300 will be evaluated for an experimental condition leading to an OSNR* ideally measured by a noise-free OSA (0.1 nm bandwidth) to be equal to 30 dB. The device 300 allows one to reconstruct the mean <ri> and the standard deviation <Δn> of the photon number distribution with single photon resolution. From <n> and <Δn> it is possible to obtain the OSNR defined by (5) and furthermore, for optically amplified signals, it is also possible to obtain OSNR* in the manner described hereafter. The mean <n> and the variance <Δn > can be expressed as:

< n >=< n_s > + < n_ASE > (6a) < An² >=< n_s >+ < n_ASE >+ < n_ASE >² +2 < n_s > ^■ < «_ASE > (6b), wherein <n$> is the mean photon number of the signal and <R_ASE> is the mean photon number of a single mode of ASE. The photon number variance expressed by eq. (6b) is the sum of four contributions: respectively the signal shot noise, the ASE shot noise, the ASE-ASE beating and the signal-ASE beating. Typically <«_ASE ^> is much less than <ns>, so the eqs. (6a-b) can be well approximated by

< n >≤< n_s > (7a) < Δn² >≡< n_s > +2 < n_s > ■ < n_ASE > (7b)

In order to extract the information about the ASE power with a sufficient precision, the contribution to the variance of eq. (7b) due to the signal- ASE beating must be greater or equal to about one. So the sensitivity of the OSNR measurement made by the photon statistics reconstruction device 300 can be determined by imposing the condition

2 < /J_s > - < «_ASE >= l (8).

The mean photon number <«_ASE> of a single mode of ASE, derived from the photon statistics reconstruction operated by the device 300, is related to the ASE power P_ASE in a reference bandwidth B_opt, which is the denominator in the definition of OSNR*, by the following relation:

^ π -^- P ASE _ ¹ P ASE

^ASE M- B_el - hv B_0≠ - hv wherein M = B_Qpil B_e\ is the number of ASE modes, B_t\ is the electrical bandwidth of the receiver and hv (Planck constant by optical frequency) is the single photon energy. In correspondence of an OSNR* equal to 30 dB, it results P_ASE = Ps / 1000. Moreover the mean photon number <ns> of the signal is related to the signal power Ps by the following relation

^s B_Λ - hv

Hence the eq. (8) can be expressed as

and for the exemplarily case of B_opt = 12.5 GHz (corresponding to 0.1 nm), B_e\ = 2 MHz and optical signal at 1550 nm wavelength, a sensitivity Ps of about - 63.5 dBm is obtained. This is much less than the power required by previously reported techniques. For example an optical spectrum analyzer usually requires tapped optical signal average power greater than about lμW (i.e. greater than -3OdBm) in order to measure OSNR* (0.1 nm bandwidth) of about 30 dB.

Preferably, the portion of optical signal S_IN inputting the device 200 is made to pass, before the optical coupler 4, through the optical filter 15 and/or the polarization controller 13 and/or the VOA 14 and/or the isolator 11'. Preferably, the optical radiation emitted by the local oscillator 1 is made to pass, before the optical coupler 4, through the polarization controller 12 and/or the isolator 11. One of the two polarization controllers 12 or 13 serves to advantageously guarantee that the polarizations of the optical signal under monitoring and of the radiation from the local oscillator are parallel in the optical coupler 4. This is advantageous in that the apparatus 300 measures the quadrature distribution of the signal component polarized parallel to the local oscillator. The other of the two polarization controllers 13 or 12 serves to advantageously control the overall balance of the device 300 which is sensible to the polarization dependent behavior of the coupler 4. hi other words, when the polarization of the optical radiations in the optical coupler deviates from the optimal 50-50 working point, the two outputs from the coupler 4 are unbalanced, causing an overall unbalancing of the device 300. In case the fluctuations of the signal polarization in the optical line 130 of Figure 1 are slower than the photon statistics reconstruction speed, it is sufficient to adjust (for example manually) the polarization controller 12 and/or 13 so as to properly align the polarizations of the local oscillator and of the optical signal in the coupler 4 only once before the measure. Typically, there is no need to adjust again the polarization controller 12 and/or 13 during the statistics reconstruction time, which typically takes few tens of seconds compared to polarization fluctuations in a time scale of several minutes. Otherwise, if the polarization fluctuations along the line are faster than the reconstruction speed, either the measure may be repeated or automatic polarization controller(s) may be used as the polarization controller 12 or/and 13.

The frequency locking functionality is achieved by the remaining part of the setup, hi operation, the portion of optical radiation emitted by the local oscillator 1 extracted by the coupler 2 and the portion of the input optical signal S_IN extracted by the coupler 3 are made to interfere in the coupler 16. Preferably, the extracted portion of optical signal is made to pass, before the optical coupler 16, through the isolator 17'. Preferably, the extracted portion of optical radiation from the local oscillator 1 is made to pass, before the optical coupler 16, through the polarization controller 18 and/or the isolator 17. The polarization controller 18 may alternatively or also act on the extracted optical signal and serves to advantageously guarantee that the two polarizations are parallel in the optical coupler 16.

The mixed optical radiation outputting the optical coupler 16 is detected by the photodetector 19 which in turn emits an electrical signal which is amplified and transformed by the voltage comparator 20 into a square wave at the same frequency of the original electrical signal. Then the frequency divider 21 divides the signal frequency so that it falls into the bandwidth of the FVC 22. The FVC output is compared by comparator 23 with a reference voltage V_ref coming from the block 8 through the frequency divider 25 and the FVC 26, to generate an error signal. The error signal is integrated in 24, attenuated by VEA 28 and used to control the optical frequency of the laser source 1, which may exemplarily be done by controlling its pumping current or its cavity length. The accuracy of the frequency locking is advantageously of the order of the greater between the linewidths of the optical signal and of the local oscillator. No phase- locking between optical signal and local oscillator is advantageously needed. The gain of the frequency locking loop may be given by

Gi₀₀₀(J) = ^{KFVC VCO} , wherein K_FVC is the gain (in YlEz) of the frequency to

voltage converter 22, M is the division factor of the frequency divider 21, K_//i2τf the integrator 24 transfer function, K_A the attenuation by VEA 28, and K_Vco=YC, where 7 is the trans-conductance (in A/V) of the current driver (not shown) of the laser source 1 and C is the frequency change per unit current change (in Hz/A) of the laser source 1 (typically Kvco=500 MRzN). Advantageously, the offset frequency between the local oscillator 1 and the optical signal S_IN is locked to the frequency of the reference signal generated by the RF signal generator 8, which means that M'K'_FVC = MK_FVC , wherein

M' is the division factor of the frequency divider 25 and K'_FVC the gain (in V/Hz) of the FVC 26.

Now, the sensitivity of the frequency locking device 400 will be evaluated. The voltage signal V read by the photodiode 19 is given by:

V'= 2G^P_sP_LO sm(2π(f_s - f_L0)t + φ) + Gn(O where G is the trans-impedance gain (in V/W) of the amplifier comprised in the photodetector 19, Pw and P_s are the powers of the local oscillator and of the signal, respectively, f_s and fw are the signal and local oscillator frequencies and n(t) is the equivalent noise at the input of the photodiode 19 (which is given by the electronic noise plus the local oscillator optical noise). The latter term must be smaller than the beat term between the signal and the local oscillator, so that:

wherein

is the total input noise variance. Exemplarily, for a noise equivalent power (NEP) at the input of the photoreceiver 19, constituted by a photodiode followed by a transinipedance amplifier, equal to 25 pW/VHz and a noise bandwidth of 100 MHz, it results σ^=6.25*10^"14 W². Assuming exemplarily a local oscillator power at the first input of the coupler 16 of about 1 mW, the required minimum signal power at the second input of the coupler 16 turns out to be equal to about 15 pW (i.e. -78 dBm), which is typically lower than the sensitivity of the photon statistics reconstruction device 300.

An advantage of this first embodiment of the device 400 is the possibility to use constitutional elements easily available on the market. Figure 3 shows a schematic diagram of a monitoring apparatus 200' in accordance with an alternative exemplary realization of the present invention. When applicable, reference to the same elements shown in Figure 2 is made using the same reference numerals of Figure 2.

The monitoring apparatus 200' shown in Figure 3 differs from that shown in Figure 2 only in the device 400' used to lock the frequency of the local oscillator 1. Accordingly, for the description of the structure and operation of the device 300 of Figure 3 and of that parts of the locking device 400' which do not differ in structure and operation from those described above, including the optical couplers 2 and 3, reference is made to the corresponding description of Figure 2. The locking device 400' differs from that (400) shown in Figure 2 in that a phase comparator 34 is downstream connected to a prescaler 31 and to a prescaler 35. A PID controller 38 (i.e proportional-integrative-derivative) is further downstream connected to the phase comparator 34 and operatively connected to the local oscillator 1 (e.g. via a driver, not shown). hi operation, the prescaler 31 divides the signal frequency of the measurement signal, which is then made to input a phase comparator 34 in which the signal is compared with the sinusoidal reference coming from the RF signal generator 8 through the prescaler 35. The output of the phase comparator 34 is a ramp signal with slope proportional to the difference frequency between the two inputs to the comparator. This output is sent to the PID controller 38 and then to the laser source current control (not shown). The loop forces the two inputs of the phase comparator 34 to have the same frequency so that the offset frequency of the local oscillator 1 and of the optical signal under monitoring becomes equal to the frequency of the signal generated by the RF signal generator 8. The phase comparator 34 preferably have a large dynamic (typically between -2-κ x 2 ^π and 2π x 2 ^π), otherwise the phase fluctuations would disturb the frequency locking operation. An advantage of this second embodiment is the possibility to use the sinusoidal reference directly, while in the embodiment of Figure 2 conversion with another frequency to voltage converter was necessary to generate a reference voltage. hi a further embodiment of the present invention, the optical monitoring apparatus 200 may monitor the OSNR and/or the average optical power of a plurality of optical channels belonging to a WDM spectrum. In this case the local oscillator 1 has the additional capability of being wavelength tunable so that, in operation, the OSNRs and/or the average powers of the desired plurality of optical channels are measured in temporal succession, by way of a preliminary wavelength tuning of the local oscillator 1 to the desired optical channel.

Although the present invention has been disclosed and described by way of some embodiments, it is apparent to those skilled in the art that several modifications to the described embodiments, as well as other embodiments of the present invention are possible without departing from the spirit or essential features thereof/the scope thereof as defined in the appended claims.

Claims

1. An optical communication system (100) comprising an optical-signal-to-noise- ratio monitoring apparatus (200) including: - a local oscillator (1);

- a 2x2 optical coupler (4) adapted to mix an optical radiation from the local oscillator (1) with a portion of an optical signal propagating through the optical communication system (100), said optical coupler (4) having two output ports adapted to emit two mixed optical radiations; - a balanced detector (5) optically coupled to said two output ports, adapted to receive said two mixed optical radiations and to emit a measurement signal;

-a calculating device (10) adapted to derive the optical-signal-to-noise-ratio of said optical signal from said measurement signal; characterized in that -the local oscillator (1) is adapted to emit a continuous wave optical radiation; and

-the optical-signal-to-noise-ratio monitoring apparatus (200) further includes a frequency locking device (400) coupled to the local oscillator (1) and adapted to lock the optical frequency of the local oscillator (1) as a function of the optical frequency of said optical signal.

2. The optical communication system (100) of claim 1 wherein the calculating device (10) is adapted to derive said optical-signal-to-noise ratio via an algorithm suitable to derive, from said measurement signal, statistical information related to the photon number probability distribution of said portion of optical signal.

3. The optical communication system (100) of claim 2 wherein said statistical information comprises the photon number probability distribution of said portion of optical signal.

4. The optical communication system (100) of claim 2 wherein said statistical information comprises statistical moments of the photon number probability distribution of said portion of optical signal.

5. The optical communication system (100) of claim 4 wherein said statistical moments comprise the mean and the variance of the photon number probability distribution of said portion of optical signal.

6. The optical communication system (100) of claim 1 or 2 wherein the measurement signal contains information on the optical signal quadrature in-phase with the optical radiation from the local oscillator.

7. The optical communication system (100) of claim 6 wherein the calculating device (10) is adapted to reconstruct the phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

8. The optical communication system (100) of claim 7 wherein the calculating device (10) is adapted to derive said optical-signal-to-noise ratio via an algorithm suitable to derive the photon number probability distribution of said portion of optical signal via a pattern functions weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

9. The optical communication system (100) of claim 7 wherein the calculating device (10) is adapted to derive said optical-signal-to-noise ratio via an algorithm suitable to derive the statistical moments of the photon number probability distribution of said portion of optical signal via a Hermite polynomial weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

10. The optical communication system (100) of claim 1 further comprising an optical tapping device (210) configured to extract said portion of the optical signal propagating through the optical communication system (100) and to input said portion to the optical coupler (4).

11. A method for optical communication comprising monitoring the optical-signal- to-noise-ratio of an optical signal according to the following steps:

- tapping a portion of said optical signal;

- mixing said portion of optical signal with an optical radiation;

- deriving the optical-signal-to-noise-ratio of said optical signal from said measurement signal; characterized in that

- said optical radiation is a continuous wave; and

- the method further comprises the step of locking the optical frequency of said optical radiation as a function of the optical frequency of the optical signal.

12. The method of claim 11 wherein said optical-signal-to-noise-ratio is derived using an algorithm suitable to derive, from said measurement signal, statistical information related to the photon number probability distribution of said portion of optical signal.

13. The method of claim 12 wherein said statistical information comprises the photon number probability distribution of said portion of optical signal.

14. The method of claim 12 wherein said statistical information comprises statistical moments of the photon number probability distribution of said portion of optical signal.

15. The method of claim 14 wherein said statistical moments comprise the mean and the variance of the photon number probability distribution of said portion of optical signal.

16. The method of any of claims 11 to 15 wherein the measurement signal contains information on the quadrature of said portion of optical signal in-phase with said optical radiation.

17. The method of claim 16 wherein deriving the optical-signal-to-noise-ratio comprises reconstructing the phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation.

18. An apparatus (200) for monitoring the optical-signal-to-noise-ratio of an optical signal, the apparatus including:

- a local oscillator (1);

- a 2x2 optical coupler (4) having a first input port optically coupled to the local oscillator (1) and a second input port adapted to receive the optical signal, the optical coupler (4) being adapted to mix an optical radiation from the local oscillator (1) with the optical signal and further having two output ports adapted to emit the mixed optical radiation in two portions;

- a balanced detector (5) optically coupled to said two output ports, adapted to receive said two portions of the mixed optical radiation and to emit a measurement signal;

-a calculating device (10) adapted to derive the optical-signal-to-noise-ratio of said optical signal from said measurement signal; characterized in that

-the local oscillator (1) is adapted to emit a continuous wave optical radiation; and

-the apparatus (200) further includes a frequency locking device (400) coupled to the local oscillator (1) and adapted to lock the optical frequency of the local oscillator (1) as a function of the optical frequency of the optical signal.

19. The apparatus (200) of claim 18 wherein the calculating device (10) is adapted to derive said optical-signal-to-noise-ratio via an algorithm suitable to derive, from said measurement signal, statistical information related to the photon number probability distribution of said portion of optical signal.

20. The apparatus (200) of claim 19 wherein said statistical information comprises the photon number probability distribution of said portion of optical signal.

21. The apparatus (200) of claim 19 wherein said statistical information comprises statistical moments of the photon number probability distribution of said portion of optical signal.

22. The apparatus (200) of claim 21 wherein said statistical moments comprise the mean and the variance of the photon number probability distribution of said portion of optical signal.

23. The apparatus (200) of any of claims 18 to 22 wherein the measurement signal contains information on the optical signal quadrature in-phase with the optical radiation from the local oscillator.

24. The apparatus (200) of claim 23 wherein the calculating device (10) is adapted to reconstruct the phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

25. The apparatus (200) of claim 24 wherein the calculating device (10) is adapted to derive said optical-signal-to-noise ratio via an algorithm suitable to derive the photon number probability distribution of said portion of optical signal via a pattern functions weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.

26. The apparatus (200) of claim 24 wherein the calculating device (10) is adapted to derive said optical-signal-to-noise ratio via an algorithm suitable to derive the statistical moments of the photon number probability distribution of said portion of optical signal via a Hermite polynomial weighted average of said phase-averaged statistical distribution of the optical signal quadrature in-phase with the optical radiation from the local oscillator.