Cross-Connect for an Optical Network

This invention relates to a cross-connect for an optical network.

In order to cope with the increasing demand for subscriber services on telecommunication networks, digital optical transmission networks are being introduced in which local access networks, which typically route electrical signals, are interconnected by a fibre optic network. This has been facilitated by the introduction of the Synchronous Digital Hierarchy (SDH) standard that deals with the formatting of signals in Synchronous Transfer Modules (STM) which will allow a unified telecommunication infrastructure and provides improved flexibility by permitting electrical digital signals from the local access networks to be converted into optical signals and transmitted through the optical network.

In order to upgrade the capacity of the optical network, Wavelength-Division-Multiplexing (WDM) techniques have been proposed, which permit the transmission capacity of a fibre link to be upgraded to the multi-Gbit/s range.

Thus, it has been proposed to provide a multi-Gbit/s WDM network superimposed on the top level of a SDH network. For a fuller discussion, reference is directed to I. Hawker, "Evolution of Digital Optical Transmission Networks", BT Technol. J., Vol. 9, No. 4, October 1991, pp 43-5. It is proposed to include nodes in the WDM network to allow dynamic re-routing of WDM data streams. It is proposed that each node should include access ports to allow optical streams to be dropped to and added from lower levels of the network. It has been proposed that routing of the high bit rate streams is directly performed in the optical layer whereas processing of any of the streams is achieved in the electronic domain by dropping it to a lower layer of the network. Network management and administration is proposed to be achieved from a management unit which configures an optical cross-connect at each node according to traffic requirements. Thus, optical cross-connects (OXC's) constitute the routing nodes of the WDM transport network.

A number of prior proposals have been made for the node architecture of the OXC and an example is given in FIG. 1.

Three input fibres 1, 2, 3 are connected at the node to output fibres 4, 5, 6. Each of the fibres 1-3 carries four WDM wavelength channels, which are connected to the output fibres 4-6 by means of space switches X.sub.1 -X.sub.4. In this example, the space switches are 4.times.4 matrices which thus have four inputs a-d and four outputs e-h that are optically interconnectable and are controlled by applied electrical signals (not shown). Such space switches are well known and include electrically controllable elements of variable refractive index to produce selective switching between the inputs and outputs.

The four optical channels of each input fibre 1-3 are applied to an input of each of the space switches X.sub.1 -X.sub.4 and each is separately selected by means of a respective tunable filter F1.sub.1, F2.sub.1, F3.sub.1, F4.sub.1. Similarly, the various wavelength channels from fibres 2 and 3 are applied to respective inputs of the space switches through filters F1.sub.2 -F4.sub.2, F1.sub.3 -F4.sub.3. The fourth input of each space switch X.sub.1 -X.sub.4 is connected to tunable optical transmitter units 7 which produce optical signals of a single frequency in response to digital electrical signal streams from the lower electrical layer of the network, to allow electrical signals to be added into the optical network for transmission in the optical layer. The electrical signals may be SDH, PDH, ATM, X25 or any other suitable format, since the optical network can transmit the bit stream transparently in the optical domain. Similarly the outputs h of each of the space switches X.sub.1 -X.sub.4 are connected to receiver units 8 that include photodetectors for producing electrical signals, so that digital data streams can be dropped out of the optical network into the lower, electrical layer. The outputs e-g of the space switches X.sub.1 -X.sub.4 are respectively connected to the inputs to the fibres 4, 5 and 6 by means of combiners 9, 10, 11. Thus, by applying control signals to the space switches X.sub.1 -X.sub.4 traffic in the WDM channels of input fibres 1-3 can be switched selectively between output fibres 4-6 and signal trains can be added to and dropped out of the optical network by means of the transmitters and receivers 7, 8. This arrangement however suffers from a number of disadvantages.

Firstly, the ultimate size of the OXC is limited by the size of the space switches. Currently, these are made in LiNbO.sub.3 or InP and are limited for practical purposes to 8.times.8. Also, for polarisation independent operation they typically require high switching voltages of the order of +/-100 V and have long rise times >100 ns. Furthermore, the devices are not at present hermetically sealed, and the material thereof tends to exhibit drift characteristics, so as to alter the performance with time.

Secondly, a node configuration as shown in FIG. 1 produces a loss of approximately 30 dB and the main contributor to the loss is the switch array. In order to overcome this problem, optical fibre amplifiers such as erbium doped fibre amplifiers need to be used to provide amplification.

Thirdly, the node architecture exhibits blocking characteristics. It is not possible for the same wavelength channel from two of the input fibres to leave on the same output fibre.

The invention provides an improved cross-connect, suitable for an optical network, which may overcome at least some of these problems. In accordance with the invention there is provided an optical cross-connect comprising: a plurality of input means each for receiving a respective group of input signals and deriving in response to each signal of the group an individual optical signal of a respective given wavelength, a plurality of output means each including a plurality of receptors selectively responsive to respective ones of the wavelengths of said optical signals to provide an output signal for the output means, and a plurality of coupling means each for coupling a respective one of said optical signals, each of a different wavelength, from each of the input means, to a respective receptor of each of the output means.

In accordance with the invention, it is possible to route signals of the same wavelength from different input means co a particular one of the output means. Furthermore, the losses involved may be substantially reduced.

In one form, the plurality of coupling means comprises a plurality of star couplers, and in another form comprises a plurality of optical bus lines to which the input and output means may be selectively added and removed as modules.

The input signals may be electrical or optical and the plurality of input means may be configured to process optical and electrical signals, to allow electrical signals to be added into the optical signal traffic. Similarly, at least one of the output means may produce electrical output signals to allow optical signals to be dropped to an electrical network.

Furthermore, an improved non-blocking arrangement may be provided, in which a plurality of the input means are each operative to receive a respective group (k) of input signals and derive in response to each signal of the group an individual optical signal of a respective given wavelength, the individual output signals being provided on selected ones of a plurality of output channels greater than k, a plurality of the output means each including a plurality of receptors greater in number than k, are selectively responsive to respective ones of the wavelengths of said optical signals to provide an output signal for the output means, and a plurality of the coupling means greater in number than k, are each operative to couple a respective one of said optical signals, each of a different wavelength, from each of the input means, to a respective receptor of each of the output means.

In order that the invention may be more fully understood an embodiment thereof will now be described by way of example, reference being had to the accompanying drawings in which:

FIG. 1 is a schematic block diagram of a prior OXC, described hereinbefore;

FIG. 2 is a schematic block diagram of an optical cross-connect in accordance with the invention;

FIG. 3 illustrates schematically input and output modules of the cross-connect shown in FIG. 2;

FIG. 4 illustrates signal blocking in the arrangement of FIGS. 2 and 3;

FIG. 5 is a schematic block diagram of the cross-connect with an alternative form of input module which avoids signal blocking;

FIG. 6 is a schematic block diagram of an alternative form of cross-connect in accordance with the invention, which utilises a back plane bus;

FIG. 7 illustrates a further optical cross-connect in accordance with the invention; and

FIG. 8 illustrates schematically input and output modules of the cross-connect shown in FIG. 7.

Referring to FIG. 2, the OXC in accordance with the invention is configured for use at a node in a WDM hierarchical telecommunication system that includes an optical layer overlying various electronic layers.

The OXC includes a number of input means in the form of input modules 10.sub.1 -10.sub.n to which optical fibres 11.sub.1 -11.sub.n are connected, the optical fibres carrying k channel WDX signals. The OXC also receives inputs from the electronic layer that are applied to input modules 12.sub.1 -12.sub.n '. Each input module 12 receives k electrical lines 13 carrying electrical input digital signals.

As will be explained in more detail hereinafter, signals corresponding to the k channels of each electrical and optical input, are connected by means of k star couplers 14.sub.1-14.sub.k to a plurality of output means in the form of optical output modules 15.sub.1 -15.sub.n and electronic output modules 16.sub.1 -16.sub.n '. The output modules 15.sub.1 -15.sub.n provide k channel WDM optical signals on output fibres 17.sub.1 -17.sub.n. The output modules 16 provide electrical outputs that can be dropped down to the electronic layer of the network on lines 18.sub.1-18.sub.n'.

The structure of a first example of input module 10 is shown in FIG. 3A and consists of a demultiplexer 19 which demultiplexes k channel WDM signals on input fibre 11 onto k parallel channels, each of which includes a wavelength converter 20.sub.1 -20.sub.k. The wavelength converters each produce a respective different wavelength in response to the input signal applied thereto so that the device shown in FIG. 3A provides one of N different output wavelengths to be coupled to the star couplers 14 as will be described hereinafter, where N corresponds to the total number of input or output modules (n+n'). The wavelength converters may comprise all-optical devices or may include an opto-electrical detector that drives a tunable laser. Examples of specific wavelength converters currently available are listed in "Photonic terabit/s networks and their key components" H. R. van As, IBM Research Div (Ruschlikon) 11.sup.th Annual Conference on European Fibre Optic Communications and Networks (EFOC & N '93) Jun. 30-July 1993 pp 13-20, and in particular on page 18.

The structure of each input module 12 for electrical signals is shown in more detail in FIG. 3B. Each input module 12 receives k parallel electrical lines 13.sub.1 -13.sub.k that are connected to respective lasers 21.sub.1 -21.sub.k that are tunable over the N different output wavelengths. Thus the lasers, which are typically laser diodes, produce output modulated signals at k different wavelengths, for the k input electrical signal channels. Specific examples of commercially available tunable lasers suitable for the system are listed on page 17 of the EFOC & N '93 paper, supra.

Referring to FIG. 2, the star couplers 14.sub.1 -14.sub.k each include N optical inputs and N optical outputs. The N inputs of each star coupler 14 are connected to receive individual different frequency outputs from each of the input modules. Thus, referring to FIG. 2, star coupler 14.sub.1 has inputs connected to receive a respective different wavelength signal train from each of the input modules 10.sub.1 -10.sub.n and 12.sub.1 -12.sub.n '. In order to simplify the drawing, only the connections from input module 10.sub.1 and 12.sub.n ' are shown. Similarly, for the star coupler 14.sub.k, connections from input module 10.sub.1 and 12.sub.n ' are shown with the other connections being omitted to simplify the drawing. The connections to star couplers 14.sub.2 -14.sub.k-1 have also been omitted for simplicity. For each of the star couplers 14, the inputs are so arranged that each receives an individual wavelength from the groups of different wavelengths produced by each of the input modules 10 and 12.

Each star coupler 14 operates in a manner well known per se so as to mix all the input signal trains and feed them to all of its outputs. Thus, considering for example star coupler 14.sub.1, a mixture of all of its inputs is applied to N outputs, namely to each of the output modules 15.sub.1 -15.sub.n, 16.sub.1 -16.sub.n '.

The nature of each of the output modules will now be described with reference to FIG. 3. Referring to FIG. 3C, the structure of each optical output module 15 is shown. The module consists of k inputs 22.sub.1 -22.sub.k which are individually connected to one output of each of the star couplers 14.sub.1 -14.sub.k. Each input 22 is connected to a receptor in the form of a tunable filter 23.sub.1 -23.sub.k which is individually tunable in order to select individual wavelengths from the mixture of signals applied to each input 22. Commercially available tunable filters are described in detail on page 17 of the EFOC & N '93 paper, supra, and may comprise electro-mechanically tuned filters (e.g. Fabry-Perot etalon), acousto-optical or semiconductor filters.

It will be recalled that each star coupler, connected to a corresponding one of the inputs 22, supplies a mixture of signals of different wavelengths selected from one of the outputs of each of the input modules 10, 12. Thus, each tunable filter 23.sub.1 -23.sub.k can be used to select an individual signal from one of the input modules on the basis of a wavelength selection. The tunable filters may be tuned individually by the application of an external control voltage from a control system (not shown). The output of each filter 23.sub.1 -23.sub.k is applied to a corresponding wavelength converter 24.sub.1 -24.sub.k. The wavelength converters produce signals that are combined by means of a passive combiner 25 to produce a k channel WDM signal on output fibre 17.

Thus, by suitable selection of the pass frequencies of the filters 23, selection of individual signals can be achieved for each output fibre 17 so that a selective cross-connection can be achieved between the input fibres 11 and output fibres 17. Moreover, electrical signals applied to input modules 12 can be added into the k channel WDM signals produced at output fibres 17, thus lifting the electrical signal from the electronic layer of the network to the optical layer.

FIG. 3D shows a structure for each electrical module 16, which allows optical signal trains to drop down from the optical layer to the electrical layer of the network. Each output module 16 consists of k input lines 26.sub.1 -26.sub.k which are applied to tunable filters 27.sub.1 -27.sub.k that function in a similar manner to the tunable filters 23 so as to select individual signal wavelengths. The outputs of the tunable filters 27 are applied to corresponding receivers 28.sub.1 -28.sub.k, typically in the form of photodiodes, which produce electrical signals corresponding to the signals of selected wavelength. The resulting k digital signal trains are applied to k output lines, which make up each output 18. Thus, signals of particular wavelengths are selected by the tunable filters 27 and are dropped down to electrical signals for use in the electronic layer of the network.

A problem with the configuration described so far is that signal blocking can occur at some of the outputs 17. This can be understood from FIG. 4 which is an enlarged, partial illustration of the configuration of FIG. 2, FIG. 3A and FIG. 3B, showing input modules 10.sub.1, 10.sub.2, star coupler 14.sub.1 and output module 15.sub.1, the other components having been omitted to simplify the drawing.

Referring to input module 10.sub.1, the input WDM signal is demultiplexed by demultiplexer 19.sub.1 to provide k signals of wavelength .lambda..sub.1 -.lambda..sub.k which are applied to respective wavelength converters 20.sub.1l -20.sub.1k. The wavelength converters produce respective outputs .lambda..sub.i -.lambda..sub.m.

A similar set of signals is produced by input module 10.sub.2, comprising wavelengths .lambda..sub.j '.lambda..sub.n as shown in FIG. 4. Each of the other input modules (not shown) produce similar outputs. Considering now the star coupler 14.sub.1, each of its N inputs receives a signal from the first wavelength converter 20.sub.11 -20.sub.21, etc. and the arrangement is set up so that each of the input wavelengths is different. Thus N separate wavelengths are applied to the star coupler. Accordingly, referring to FIG. 4, .lambda..sub.i .noteq..lambda..sub.j. As previously explained, the resulting mix of signals produced by star coupler 14.sub.1 is applied to a single input of each of the output modules 15.sub.1 -15.sub.n, 16.sub.1 -16.sub.n'. Thus, considering the output module 15.sub.1 as shown in FIG. 4, it is not possible to select wavelength .lambda..sub.i and .lambda..sub.j from the input modules 10.sub.1, 10.sub.2 simultaneously at the output module 15.sub.1 since both of the signals are applied simultaneously to one of the filters 23. In FIG. 4 this is shown as input filter 23.sub.1. The filter 23.sub.1 can select only one of the wavelengths .lambda..sub.i, .lambda..sub.j and therefore, signal blocking of one of the wavelengths will occur.

This problem is overcome by the configuration shown in FIG. 5, in which a different form of the input module 10' is provided. Referring to input module 10'.sub.1 the input WDM signals applied on input fibre 11.sub.1 are fed to a beam splitter 30.sub.1 that provides k parallel channel outputs each of which is fed to a tunable filter 31.sub.1l -31.sub.1k that in turn are connected to individual wavelength converters 20.sub.1l -20.sub.1k. Thus, by tuning the individual filters 31.sub.1l -31.sub.1k, the different input signals .lambda..sub.1 -.lambda..sub.k can be applied to different ones of the k channels selectively. The wavelength converters 20.sub.1l -20.sub.1k then produce appropriate signals for the inputs of the star couplers 14. In FIG. 5, the first two star couplers 14.sub.1, 14.sub.2 are shown. When it is desired to assemble a signal at output module 15.sub.1 that contains wavelength .lambda..sub.i from input module 10'.sub.1 and .lambda..sub.j from input module 10'.sub.2, if, as previously described with reference to FIG. 4, these signals were applied to the inputs of a single one of the star couplers, blocking would occur. However, as shown in FIG. 5, the tunable filter 31.sub.22 is adjusted such that wavelength converter 20.sub.22 (rather than converter 20.sub.21) produces the output wavelength .lambda..sub.j. Consequently, the signal of wavelength .lambda..sub.j is applied to star couplers 14.sub.2 and hence to the second input of the output module 15.sub.1. Thus, the signals of wavelength .lambda..sub.i and .lambda..sub.j can be selected by the tunable filters 23.sub.1 and 23.sub.2 respectively in the output module 15.sub.1. In this way, signal blocking is avoided.

Referring to FIG. 6, this shows an alternative configuration in which the star couplers of FIG. 2 are replaced by a back plane bus. Thus, the individual input modules 10.sub.1 -10.sub.n, 12.sub.1 -12.sub.n have their k outputs connected individually to k parallel optical fibres 32 which form a fibre back plane. The output modules 15.sub.1 -15.sub.n and 16.sub.1 -16.sub.n have their respective k inputs selectively connected to the k fibres of the back plane. This arrangement is particularly suited to use with input modules of the form shown in FIG. 5 since individual signal trains can be programmably applied to individual ones of the back plane fibres 32 and then selectively received by the output modules 15, 16 under the control of tunable filters 23, 27. This arrangement has the advantage that the individual modules can be arranged as units that can be plugged onto the back plane for example by using optical D-couplers.

The described examples of the invention have the advantage that the number of elements that need to be controlled is much smaller as compared with the configuration shown in FIG. 1. Considering an example of the OXC of FIG. 2, when compared with that of FIG. 1, when the number of wavelengths k=4 and the number of input and output optical lines is 3, the component count is set out in the following table:

Furthermore, since the wavelength converters in the input modules can be permanently set such that the output wavelength of a converter is always the same, the number of elements requiring control can be reduced to 36.

The optical star couplers of FIG. 2 and the back plane bus of FIG. 6 are inexpensive components as compared with the space switches shown in FIG. 1. Packaged optical star couplers which are wavelength insensitive between 1300-1600 nm and as large as 144.times.144 have been reported and are thus suited to implement the invention. Since the optical star couplers are themselves inexpensive compared to the input and output modules shown in FIG. 2, it is envisaged that not all the input and output modules need be attached at the time of commissioning the system so that the system can readily be upgraded in terms of its capacity thereafter. The described input and output modules both include wavelength converters, with the advantage that if a signal inversion occurs at the wavelength converter, two inversions occur, so that no signal inversion will occur in the final output of the cross-connect.

An embodiment of the invention will now be described in A which an alternative way of avoiding channel blocking is achieved. Referring to FIG. 7, the structure is generally similar to that shown in FIG. 2 and like parts are given the same reference numbers. Each of the input optical fibres 11 carries k optical multiplex channels, but each input module 10, instead of providing k parallel channels, provides Q channels where:

Similarly, the input modules 12 provide Q parallel channels and Q star couplers 14 are provided. Each of the output modules 15, 16 is provided with Q parallel channels.

The various input and output modules 10, 12, 15 and 16 are shown in more detail in FIGS. 8a-d respectively. Referring to FIG. 8a, each input module 10 is similar to that shown in FIG. 5 but with Q channels, and consists of an input fibre 11 that carries k input multiplexed channels, connected to a passive splitter 19 that has Q outputs coupled to Q tunable filters 31.sub.1 -31.sub.Q which are respectively connected to wavelength converters 20.sub.1 -20.sub.Q, in order to provide Q parallel outputs. From the foregoing, it will be seen that the channel redundancy of (Q-k) and the tunable nature of the various channels Q allows a strictly non-blocking configuration to be provided.

Referring now to FIG. 8b which shows the input module 12, the arrangement has Q input channels connected to respective tunable lasers 21.sub.1 -21.sub.Q. For a system which provides k input channels, the remaining input channels (Q-k) can be used to provide test signals. The configuration of the output modules 15, 16 shown in FIGS. 8c, d is the same as that shown in FIGS. 3c, d except for the provision of Q parallel channels rather than k. It will be seen that by appropriate selection of the wavelength of operation of tunable filters 31 and 23, 27, a fully non-blocking architecture is provided by virtue of the (Q-k) channel redundancy. Whilst this non-blocking arrangement has been described for use with the star couplers 14, it will be appreciated that it can also be used in connection with a back plane bus as shown in FIG. 6, but utilising Q back plane channels. Thus the various modules of FIG. 8, could be used with a Q channel back plane bus.

Since the wavelengths used in the cross-connect according to the invention are selected independently of the wavelengths used in the optical network itself, it is not necessary to control the preciseness of wavelength throughout the entire network but only on a link-by-link basis, since each optical cross-connect in accordance with the invention provides a decoupling between adjacent sections of the network.

Also, the separation of the N wavelengths used in the optical cross-connect need not be the same as the separation of the wavelengths .lambda..sub.1 -.lambda..sub.k used on the link. For example, the link wavelengths .lambda..sub.1 -.lambda..sub.k may each be separated by 2 nm whereas the wavelengths .lambda..sub.i, .lambda..sub.n can be separated by 1 nm, 2 nm, 3 nm or any other suitable spacing.

Thus, the set of wavelengths used in the optical cross-connect is independent of the network and the number of trunk lines and the number of channels k used in the network itself. Thus, N can be chosen independently of k.