Distributed space-time-space switch Number:7,394,806 from the United States Patent and Trademark Office (PTO) owispatent A wide-coverage, high-capacity, switching network is modeled after a classical space-time-space switch. In the switching network, each of the space stages comprises geographically distributed optical space switches and the time stage comprises a plurality of geographically distributed high-capacity electronic switching nodes. User-access concentrators, each supporting numerous users, access the network through ports of the distributed optical space switches. A user-access concentrator is a simple device which need only have a single access channel to access the network, although two or more access channels may be used. Such a user-access concentrator can communicate with a large number of other user-access concentrators by time-multiplexing the access channel.<H Distributed, switch, Distributed, spa, 7394806.html, Patent, pto, 7394806

Title: Distributed space-time-space switch

Abstract: A wide-coverage, high-capacity, switching network is modeled after a classical space-time-space switch. In the switching network, each of the space stages comprises geographically distributed optical space switches and the time stage comprises a plurality of geographically distributed high-capacity electronic switching nodes. User-access concentrators, each supporting numerous users, access the network through ports of the distributed optical space switches. A user-access concentrator is a simple device which need only have a single access channel to access the network, although two or more access channels may be used. Such a user-access concentrator can communicate with a large number of other user-access concentrators by time-multiplexing the access channel.

Patent Number: 7,394,806 Issued on 07/01/2008 to Beshai, et al.

Inventors: Beshai; Maged E. (Stittsville, CA), Goodwill; Dominic John (Kanata, CA)
Assignee: Nortel Networks Limited (St. Laurent, Quebec, CA)

Appl. No.: 10/390,730

Filed: March 19, 2003

Current U.S. Class: 370/380 ; 370/382

Current International Class: H04L 12/50 (20060101)

References Cited [Referenced By]

U.S. Patent Documents


5144619	September 1992	Munter
5168492	December 1992	Beshai et al.
5430722	July 1995	Jacob et al.
5475679	December 1995	Munter
5623356	April 1997	Kaminow et al.
5663818	September 1997	Yamamoto et al.
5745486	April 1998	Beshai et al.
5889600	March 1999	McGuire
6118792	September 2000	Beshai
6288808	September 2001	Lee et al.
6333799	December 2001	Bala et al.
6486983	November 2002	Beshai et al.
6552833	April 2003	Liu et al.
6570872	May 2003	Beshai et al.
6738581	May 2004	Handelman
6813407	November 2004	Ramaswami et al.
6925257	August 2005	Yoo
7116862	October 2006	Islam et al.
7171072	January 2007	Beshai et al.
7212739	May 2007	Graves et al.
2002/0015551	February 2002	Tsuyama et al.
2002/0021467	February 2002	Ofek et al.
2003/0030866	February 2003	Yoo

Foreign Patent Documents


2353744	Feb., 2002	CA
0964487	Dec., 1999	EP

Other References

US. Appl. No. 10/222,223, filed Aug. 20, 2002, Maged Beshai et al. cited by other .
U.S. Appl. No. 09/671,140, filed Sep. 28, 2000, Beshai et al. cited by other .
European Search Report for 03252289.8, mailed Jun. 6, 2005. cited by other.

Primary Examiner: Vu; Huy D.
Assistant Examiner: Nguyen; Phuongchau B
Attorney, Agent or Firm: Withrow & Terranova, PLLC

Claims

What is claimed is:

1. A distributed space-time-space switch of comprising: an electronic time-switching stage including a plurality of electronic time switches; an input stage including a plurality of upstream optical space switches, where each upstream optical space switch of said plurality of upstream optical space switches receives upstream input signals on a plurality of input channels and transmits upstream output signals to at least one electronic time switch of said plurality of electronic time switches; and an output stage including a plurality of downstream optical space switches, where each downstream optical space switch of said plurality of downstream optical space switches receives downstream input signals from at least one electronic time switch of said plurality of electronic time switches and transmit downstream output signals on a plurality of output channels, wherein each of said optical space switches is associated with a controller for scheduling connections across said each of said optical space switches, and wherein each electronic time switch of the plurality of electronic time switches is operable to transmit an optical signal to a given downstream optical space switch in said output stage so that said optical signal arrives at said given downstream optical space switch at an instant of time specified by said controller associated with said, given downstream optical space switch.

2. The distributed space-time-space switch of claim 1 wherein at least one of said optical space switches comprises a star coupler and a demultiplexer.

3. The distributed space-time-space switch of claim 2 wherein said demultiplexer comprises an Arrayed Waveguide Grating device.

4. The distributed space-time-space switch of claim 2 wherein said demultiplexer comprises an Echelle grating device.

5. The distributed space-time-space switch of claim 2 wherein said demultiplexer comprises an array of thin-film filters.

6. The distributed space-time-space switch of claim 2 wherein said demultiplexer comprises an optical filter device incorporating a fiber Bragg grating.

7. The distributed space-time-space switch of claim 2 wherein said demultiplexer comprises an optical filter device incorporating a waveguide Bragg grating.

8. The distributed space-time-space switch of claim 2 wherein said star coupler includes a plurality of star-coupler input ports and at least one star-coupler output port.

9. The distributed space-time-space switch of claim 8 wherein at least one of said star-coupler input ports receives optical signals from a wavelength tunable source.

10. The distributed space-time-space switch of claim 9 wherein at least one of said star-coupler input ports receives optical signals from a wavelength converter.

11. The distributed space-time-space switch of claim 10 wherein at least one of said downstream optical space switches comprises a first demultiplexer, a spectral-translation module, a star coupler and a second demultiplexer, said spectral-translation module including a plurality of wavelength-converters.

12. The distributed space-time-space switch of claim 1 wherein said plurality of input channels are received on a single wavelength division multiplexed link.

13. The distributed space-time-space switch of claim 1 wherein said plurality of output channels are transmitted on a single wavelength division multiplexed link.

14. A switching network comprising: a plurality of electronic time switches; a wavelength router communicatively connected to each of said plurality of electronic time switches by a wavelength-division-multiplexed link, where said wavelength router routes received signals according to wavelength; and a plurality of optical space switches, wherein each optical space switch of said plurality of optical space switches is communicatively connected to said wavelength router by the wavelength-division-multiplexed link; wherein each optical space switch of said plurality of optical space switches includes: an upstream optical space switch which: receives upstream signals from a plurality of user-access concentrators; and switches said upstream signals toward said wavelength router; and a downstream optical space switch which: receives downstream signals from said wavelength router; and switches said downstream signals toward said plurality of user-access concentrators.

15. The switching network of claim 14 wherein said wavelength router comprises an upstream wavelength router and a downstream wavelength router, wherein: said upstream wavelength router communicatively connects to each of said plurality of electronic time switches by a corresponding upstream wavelength-division-multiplexed link; and said downstream wavelength router communicatively connects to each of said plurality of electronic time switches by a corresponding downstream wavelength-division-multiplexed link.

16. A switching network comprising: a plurality of electronic time switches; a plurality of wavelength routers, each of said plurality of wavelength routers connecting to each of said plurality of electronic time switches by corresponding wavelength-division-multiplexed links; a plurality of optical space switches arranged in a plurality of groups, wherein each of said plurality of groups is associated with a given wavelength router of said plurality of wavelength routers and each optical space switch of the plurality of optical space switches in each group of said plurality of groups communicatively connects to said associated given wavelength router by a wavelength-division-multiplexed link; and a plurality of concentrators arranged in a plurality of concentrator sets, where each concentrator set in said plurality of concentrator sets is associated with a given optical space switch of said plurality of optical space switches.

17. The network of claim 16 wherein each concentrator of the plurality of concentrators in a given concentrator set of the plurality of concentrator sets: receives incoming signals from a plurality of traffic sources; processes said incoming signals to form outgoing optical signals; and transmits said outgoing optical signals to said optical space switch associated with said given concentrator set.

18. The network of claim 17 wherein each concentrator of the plurality of concentrators in a given concentrator set of the plurality of concentrator sets: receives control signals from a controller of an optical space switch of said plurality of optical space switches; and selects wavelengths for optical carriers for said outgoing optical signals based on said control signals.

19. The network of claim 16 wherein said each concentrator of the plurality of concentrators: receives a time-slotted downstream optical signal from said optical space switch associated with said given concentrator set; processes said downstream optical signal to give individual signals; and transmits said individual signals to a plurality of traffic sinks.

20. The network of claim 16 wherein each wavelength router of said plurality of wavelength routers comprises an upstream wavelength router and a downstream wavelength router, wherein: said upstream wavelength router communicatively connects to each of said electronic time switches of the plurality of electronic time switches by a corresponding upstream wavelength-division-multiplexed link; and said downstream wavelength router communicatively connects to each of said electronic time switches of the plurality of electronic time switches by a corresponding downstream wavelength-division-multiplexed link.

21. The network of claim 20 wherein said upstream wavelength router comprises a first array of multiplexers interlaced with a first array of demultiplexers and said downstream wavelength router comprises a second array of multiplexers interlaced with a second array of demultiplexers.

22. The network of claim 21 wherein each multiplexer in said first array of multiplexers and each demultiplexer in said first array of demultiplexers is an Arrayed Waveguide Grating device.

23. The network of claim 21 wherein each multiplexer in said second array of multiplexers and each demultiplexer in said second array of demultiplexers is an Arrayed Waveguide Grating device.

24. The network of claim 16 wherein each optical space switch of said plurality of optical space switches comprises an upstream optical space switch and a downstream optical space switch and wherein: said upstream optical space switch: receives upstream optical signals from one of said concentrator sets of the plurality of concentrator sets; and selectively transmits said upstream optical signals to an upstream wavelength router in said associated given wavelength router; and each downstream optical, space switch: receives downstream optical signals from a downstream wavelength router in said associated given wavelength router; and selectively transmits said downstream optical signals to an associated concentrator set of the plurality of concentrator sets.

25. The network of claim 24 wherein said upstream optical space switch and said downstream optical space switch are associated with a mutual space switch controller.

26. The network of claim 25 wherein each concentrator of the plurality of concentrators in a given concentrator set transmits an outgoing optical signal to said optical space switch associated with said given concentrator set so that said outgoing optical signal arrives at said optical space switch associated with said given concentrator set at an instant of time specified by said mutual space switch controller of said optical space switch associated with said given concentrator set.

27. The network of claim 26 wherein said each concentrator of the plurality of concentrators in said given concentrator set transmits said outgoing optical signal to said optical space switch associated with said given concentrator set so that said outgoing optical signal uses a wavelength channel specified by said mutual space switch controller of said optical space switch associated with said given concentrator set.

28. The network of claim 25 wherein each concentrator of the plurality of concentrators in a given concentrator set comprises a tunable laser for altering a wavelength of said outgoing optical signal to an upstream optical space switch associated with said given concentrator set to a wavelength specified by said mutual space switch controller of said optical space switch associated with said given concentrator set.

29. The network of claim 25 wherein each of said electronic time switches of the plurality of electronic time switches transmits an optical signal to said downstream optical space switch so that said optical signal arrives at said downstream optical space switch at an instant of time specified by said mutual space switch controller of said optical space switch associated with a given concentrator set.

30. The network of claim 29 wherein said upstream optical space switch comprises a star coupler and an Arrayed Waveguide Grating demultiplexer, said star coupler having a plurality of input ports and at least one output port, a one of said at least one output ports connecting to said Arrayed Waveguide Grating demultiplexer, said plurality of input ports receiving optical signals from an associated concentrator set.

31. The network of claim 29 wherein said upstream optical space switch comprises a star coupler having a plurality of input ports and at least one output port, each of said plurality of input ports receiving an optical signal from one of said plurality of concentrators via an associated wavelength converter.

32. The network of claim 31 wherein said associated wavelength converter is controlled by said mutual space switch controller.

33. The network of claim 29 wherein said downstream optical space switch comprises: a first wavelength demultiplexer, a spectral-translation module, a star coupler; and a second wavelength demultiplexer; where said star coupler has a plurality of star coupler input ports and a single star coupler output port connecting to said second wavelength demultiplexer; and where said second wavelength demultiplexer has a plurality of wavelength demultiplexer output ports and at least one of said wavelength demultiplexer output ports communicatively connects to one of said plurality of concentrators.

34. The network of claim 33 wherein selected ones of said star coupler input ports receive an optical signal from said spectral-translation module and said spectral-translation module is controlled by said mutual space switch controller of said optical space switch associated with said given concentrator set.

35. A switching node comprising: an input array of optical space switches; a middle array of switches including optical space switches and time-space switches; an output array of optical space switches; a plurality of first switch controllers for controlling said optical space switches in said input array and said output array; a plurality of second switch controllers for controlling said optical space switches in said middle array of switches; and a plurality of third switch controllers for controlling said time-space switches in said middle array of switches, wherein: each of said optical space switches in said input array has a link to each of said optical space switches in said middle array of switches; each of said optical space switches in said input array has a link to each of said time-space switches in said middle array of switches; each of said optical space switches in said middle array switches has a link to each of said optical space switches in said output array; each of said time-space switches in said middle array of switches has a link to each of said optical space switches in said output array; and each of said links comprises at least one channel.

36. The switching node of claim 35 wherein a path from one of said optical space switches in said input array to one of said optical space switches in said output array through one of said optical space switches in said middle array of switches is a continuous channel path.

37. The switching node of claim 36 wherein a given first switch controller of said plurality of first switch controllers communicates with a given second switch controller of said plurality of second switch controllers by shifting a carrier wavelength of a channel transmitted from an optical space switch associated with said given first switch controller to an optical space switch associated with said given second switch controller.

38. The switching node of claim 35 wherein a path from one of said optical space switches in said input array to one of said optical space switches in said output array through one of said time-space switches in said middle array of switches is a time-slotted path.

39. The switching node of claim 38 wherein a given first switch controller of said plurality of first switch controllers transmits control signals to a given third switch controller of said plurality of third switch controllers during a designated time slot in a time-slotted frame.

Description

FIELD OF THE INVENTION

The present invention relates to communication networks and, more particularly, to the architecture and control of a distributed space-time-space switch and a switching network modeled on such a switch.

BACKGROUND

Network expansion is motivated by the prospects of new applications requiring a much higher capacity than that required by today's applications and is facilitated by the abundance of data transport capacity (often called bandwidth) of the optical telecommunication medium. The realizable capacity of a telecommunication network is virtually unlimited. A network structure that enables virtually unlimited expansion while providing a high service quality is desirable and its introduction is overdue.

Current communication networks, however, are complex. For example, the current Internet is complex and inefficient, with limited scalability and service capabilities: scalability relates to the ability of a network to grow to handle increasing traffic and accommodate a greater number of nodes; service capabilities relate to the ability of a network to provide flexible intelligent services and quality guarantees of various types of service. The current Internet lacks the versatility required in a growing global multi-service network, and its structure prohibits its growth without tremendous complexity and expense. This is further complicated by the unduly complex protocols that are an accumulation of patchwork performed since the Internet's inception.

Advances in optical and electronic technology have eliminated the need for complex structures and complex controls of telecommunication networks. A versatile inexpensive network scaling to a capacity that is orders of magnitude higher than the capacity of the current Internet is now realizable using simple network structures. The limitations that have led to the complexity and inefficiency of the current data networks have now been traversed. Adopting a simple network structure would enable the construction of an economical wide-coverage high-capacity high-performance network and the introduction of advanced communication services.

Applicant's U.S. patent application Ser. No. 09/286,431 filed on Apr. 6, 1999 and titled "Self-Configuring Distributed Switch ", discloses a wide-coverage network of a composite-star structure that greatly simplifies network routing and control while facilitating growth to very high capacities. The disclosed network is based on adaptive wavelength channel allocation in an optical-core comprising several core nodes. To simplify the control functions, the core nodes operate independently from each other. The network is fully meshed and the paths have adaptive capacities. A technique for overcoming optical-switching latency in such a composite-star structure is described in U.S. Pat. No. 6,486,983, titled "Agile Optical-Core Distributed Packet Switch", issued to Beshai et al. on Nov. 26, 2002.

It is well known that fine switching granularity can reduce the number of hops in a network and, hence, increase network efficiency. On the other hand, it is also recognized that some applications are better served through channel switching. Therefore, it may be beneficial to provide a network of mixed granularity. Applicant's U.S. patent application Ser. No. 09/671,140 filed on Sep. 28, 2000 and titled "Multi-grained Network" describes a network which includes edge nodes interconnected by core nodes having distinctly different granularities. The edge nodes switch multi-rate data traffic. The core may include core nodes that switch fixed-size data blocks, core nodes that switch channels or bands of channels, and core nodes that switch entire links. A core node that provides fine granularity by time sharing--for example, by switching data blocks occupying short time slots--must have a low switching latency in order to enable efficient time-sharing of wavelength channels.

The networks disclosed in the aforementioned patent applications require that each edge node have a sufficient capacity to enable direct linkage to the core nodes. Traffic sources may then access the edge nodes directly.

With the advent of fast optical switching devices, it may be desirable to relax the requirement that each edge node be of high capacity so that edge nodes of widely-varying sizes may be used while still maintaining the precious property of a small number of hops from any traffic source to any traffic sink. This would require exploring new network structures.

SUMMARY

A distributed space-time-space switch is adapted from a known space-time-space switch. The structure of the distributed space-time-space switch may be expanded to serve as a wide-coverage, high-capacity, switching network. Advantageously, the switching network may be pre-configured in a manner that allows the switching of data traffic to happen predictably according to a wavelength chosen for a carrier of the data traffic at the edge of the network. In the switching network, each of the space stages may comprise geographically distributed optical space switches and the time stage may comprise a plurality of geographically distributed high-capacity electronic switching nodes. User-access concentrators, each supporting numerous users, may access the switching network through ports of the distributed optical space switches. A user-access concentrator is a simple device which need only have a single access channel to access the network. Such a user-access concentrator may communicate with a large number of other user-access concentrators by time-multiplexing the access channel.

According to an aspect of the present invention, there is provided a distributed space-time-space switch. The distributed space-time-space switch includes an electronic time-switching stage including a plurality of electronic time switches, an input stage including a plurality of upstream optical space switches, where each upstream optical space switch of the plurality of upstream optical space switches receives upstream input signals on a plurality of input channels and transmits upstream output signals to at least one electronic time switch of the plurality of electronic time switches and an output stage including a plurality of downstream optical space switches, where each downstream optical space switch of the plurality of downstream optical space switches is adapted to receive downstream input signals from at least one electronic time switch of the plurality of electronic time switches and transmit downstream output signals on a plurality of output channels.

According to another aspect of the present invention, there is provided a switching network. The switching network includes a plurality of electronic time switches, a wavelength router communicatively connected to each of the plurality of electronic time switches by a wavelength-division-multiplexed link, where the wavelength router is adapted to route received signals according to wavelength, and a plurality of optical space switches, wherein each optical space switch of the plurality of optical space switches is communicatively connected to the wavelength router by a wavelength-division-multiplexed link. Each of the optical space switches of the plurality of optical space switches includes an upstream optical space switch adapted to receive upstream signals from a plurality of user-access concentrators and switch the upstream signals toward the wavelength router. Each of the optical space switches of the plurality of optical space switches also includes a downstream optical space switch adapted to receive downstream signals from the wavelength router and switch the downstream signals toward the plurality of user-access concentrators.

According to a further aspect of the present invention, there is provided a switching network. The switching network includes a plurality of electronic time switches, a plurality of wavelength routers, each of the wavelength routers connecting to each of the plurality of electronic time switches by corresponding wavelength-division-multiplexed links, a plurality of optical space switches arranged in a plurality of groups, wherein each of the groups is associated with a given wavelength router of the plurality of wavelength routers and each optical space switch in each group of the plurality of groups communicatively connects to the associated given wavelength router by a wavelength-division-multiplexed link and a plurality of concentrators arranged in a plurality of concentrator sets, where each concentrator set in the plurality of concentrator sets is associated with a given optical space switch of the plurality of optical space switches.

According to a still further aspect of the present invention, there is provided a switching node. The switching node includes an input array of optical space switches, a middle array of switches including optical space switches and time-space switches and an output array of optical space switches, wherein each of the optical space switches in the input array has a link to each of the optical space switches in the middle array of switches, each of the optical space switches in the input array has a link to each of the time-space switches in the middle array of switches, each of the optical space switches in the middle array of switches has a link to each of the optical space switches in the output array and each of the time-space switches in the middle array of switches has a link to each of the optical space switches in the output array.

According to an even further aspect of the present invention, there is provided a method of communicating a control signal along a channel carrying a payload signal modulating a current wavelength. The method includes shifting the current channel from the current wavelength to a prescribed control wavelength, interrupting the payload signal and causing the control wavelength to be modulated with the control signal.

According to still another aspect of the present invention, there is provided a user-access concentrator. The user-access concentrator includes a traffic interface adapted to receive upstream signals from a plurality of traffic sources, a tunable optical transmitter adapted to produce a tunable optical carrier signal modulated with the upstream signals to give outgoing optical signals, a concentrator controller adapted to control a wavelength band of the optical carrier signal and an output port adapted to transmit the outgoing optical signals to an optical space switch.

According to still another aspect of the present invention, there is provided a controller for controlling a space switch. The controller is adapted to determine a switching schedule for operation of the space switch and transmit control signals representative of the switching schedule to a plurality of network elements.

Other aspects and features of the present invention will become apparent to those of ordinary skill in the art upon review of the following description of specific embodiments of the invention in conjunction with the accompanying figures.

BRIEF DESCRIPTION OF THE DRAWINGS

In the figures which illustrate example embodiments of this invention:

FIG. 1 illustrates a known three-stage channel switch;

FIG. 2 illustrates a known time-space-time switch used for fine-granularity switching of time-slotted signals;

FIG. 3 illustrates a known space-time-space switch used for fine-granularity switching of time-slotted signals;

FIG. 4 illustrates an alternative space-time-space switch adapted from the space-time-space switch of FIG. 3;

FIG. 5 illustrates a combination switching node combining, according to an embodiment of the present invention, the features of the switches of FIG. 1, FIG. 3, and FIG. 4, where a middle stage comprises a set of channel switches and a set of baseband switches;

FIG. 6 illustrates a switch structure derived by rearranging the structure of FIG. 5.

FIG. 7 illustrates a switch structure of FIG. 6 including baseband switching modules and related connections;

FIG. 8 illustrates a network of edge nodes and bufferless core switches, used to illustrate the difficulty of time-locking paths each traversing more than one bufferless switch;

FIG. 9 illustrates an aspect of a time-locking process;

FIG. 10 illustrates the connection of user-access concentrators to a dual space switch according to an embodiment of the present invention;

FIG. 11 illustrates the structure of an exemplary one of the user-access concentrators of FIG. 10 according to an embodiment of the present invention;

FIG. 12 illustrates the structure of an optical space switch, including a space switch controller, in communication with user-access concentrators and wavelength routers according to an embodiment of the present invention;

FIG. 13 illustrates the structure of an exemplary space switch controller of FIG. 10 according to an embodiment of the present invention;

FIG. 14 illustrates a distributed space-time-space switch according to an embodiment of the present invention;

FIG. 15 illustrates an alternative arrangement of the distributed space-time-space switch of FIG. 14 according to an embodiment of the present invention;

FIG. 16 illustrates a switching network based on the distributed space-time-space switch of FIG. 15 according to an embodiment of the present invention;

FIG. 17 illustrates a fast-switching optical switch based on a single star coupler where spatial switching is effected by tunable lasers provided at the traffic sources according to an embodiment of the present invention;

FIG. 18A illustrates a fast-switching optical switch based on a single star coupler where spatial switching is effected using wavelength converters according to an embodiment of the present invention;

FIG. 18B illustrates the switch of FIG. 18A preceded by a wavelength demultiplexer for use with an embodiment of the present invention;

FIG. 19 illustrates the upstream side of a switching network comprising high-capacity baseband switches and high-capacity channel switches interconnecting fast optical switches, in accordance with an embodiment of the present invention;

FIG. 20 illustrates the downstream side of a switching network comprising high-capacity baseband switches and high-capacity channel switches interconnecting fast optical switches, in accordance with an embodiment of the present invention;

FIG. 21 illustrates time-locked upstream and downstream paths traversing a source access concentrator, an upstream space switch, a core electronic switch, a downstream space switch, and a destination. access concentrator; and

FIG. 22 illustrates the time-locked paths of FIG. 21 with each space switch implemented as a star-coupler-based space switch.

DETAILED DESCRIPTION

Before describing embodiments of the present invention, a description of the concept and realization of time-locking is provided.

A first node X is said to be time-locked to a second node Y along a given path if, at any instant of time, the reading of a time counter at node X equals the sum of a reading of an identical time counter at node Y and the propagation time, normalized to the time counter cycle duration, along the given path from node X to node Y. The time counters at nodes X and Y have the same cycle duration. There may be several paths connecting the first node to the second node, and the paths may be defined by individual wavelengths in a fiber link or several fiber links. Due to the difference in propagation delays of different paths connecting the same node pair, time-locking must be realized for the different paths individually. Due to dispersion, time-locking of individual wavelength channels within the same WDM link may be required. When a first node is time-locked to a second node along a given path, the given path is said to be time-locked. It is noted that the methods and apparatus of the present invention apply to both channel switching and TDM switching.

The time-locking process in a time-shared network is described with the help of a two-node model. To realize time-locking of a first node to a second node in a network, the first node is provided with a first controller that includes a first time counter and the second node is provided with a slave controller and a master controller that includes a master time counter. The second node has several input ports and output ports and the master controller is connected to one of the input ports and one of the output ports. The first controller sends an upstream control burst to an input port of the second node during a designated time interval, the upstream control burst including a reading of the first time counter. The upstream control burst is sent in-band, together with payload data destined to output ports of the second node. The slave controller must be able to direct the upstream control burst to the master controller during a pre-scheduled time interval. The master controller has a device for acquiring and parsing upstream control bursts. The master controller compares the reading of the first time counter with a reading of the master time counter. An agreement of the two readings, or a negligible discrepancy, ascertains time alignment. The master controller reports reading discrepancies to the first controller which resets its time counter accordingly.

Time-locking an edge node to a reference node is realized by time-locking a time counter at the edge node to a time counter at the reference node. A time counter can be a conventional clock-driven counter. A time counter at an edge node may be an up-counter and a time counter at a reference node may be a down counter, the two counters have the same cycle duration. Using a 28-bit time counter, for example, driven by a clock of a clock period of 20 nanoseconds, the duration of the time counter cycle would be about 5.37 seconds (2.sup.28 times 20 nanoseconds). The reading of an up-counter at an edge node increases, with each clock trigger, from 0 to 268,435,455 (0 to 2.sup.28-1) and the reading of a time counter at a reference node decreases, with each clock trigger, from 268,435,455 to 0. If the edge-node controller sends a timing message, when its reading is K.sub.1, to a reference node, and the reading of the down-counter of the reference node at the instant of receiving the timing message is K.sub.2, then the edge-node controller must reset its up-counter to zero when the up-counter reading reaches [K.sub.2+K.sub.1] modulo 2.sup.B, B being the wordlength of the time counter (B=28 in the above example). If K.sub.2+K.sub.1=2.sup.B-1, the edge node is already time-locked to the reference node.

Thus, within a network, all time counters have the same cycle duration and time-coordination can be realized through an exchange of time counter readings between each source node and a reference node to which the source node is connected. In a TDM (time-division multiplexing) switching network, the time counter readings may be carried in-band, alongside payload data destined to sink nodes, and sending each time counter reading must be timed to arrive at a corresponding reference node during a designated time interval.

FIG. 1 illustrates a known three stage channel switch 100 (a space-space-space switch, often abbreviated as an S-S-S switch) used for switching any input. channel from among a plurality of input channels to any output channel from among a plurality of output channels. The three array switch includes a first array 190-1, a second array 190-2 and a third array 190-3. Each of the three arrays 190-1, 190-2, 190-3 includes multiple identical space-switch modules 102.

A space-switch module is a bufferless switch that instantaneously connects any of several incoming channels to any of several outgoing channels. Space switches have graduated from electro-mechanical mechanisms with metallic contacts to electronic switches using integrated circuits, then to photonic-based switches.

Each of the space-switch modules 102 has dimension n.times.n, having n>1 input ports and n output ports, i.e., one input port and one output port for each of n space-switch modules 102. The three-stage switch 100 of FIG. 1 has a dimension of n.sup.2.times.n.sup.2, with an input capacity of n.sup.2 channels and an output capacity of n.sup.2 channels thus accommodating up to n.sup.2 concurrent connections. Several variations of the architecture of the three stage channel switch 100 of FIG. 1 can be constructed, including a folded architecture and a double-folded architecture. In the folded architecture arrangement, the first array 190-1 and the third array 190-3 of the three-stage structure are combined so that each first-array space-switch module 102 pairs with a third-array space-switch module 102 to form a combined (2.times.n).times.(2.times.n) space-switch module. In the double-folded architecture arrangement, a single array of space-switch modules may be used, with each space-switch module connecting directly to each other space-switch module to form a full mesh. In the unfolded arrangement of FIG. 1, a path from an input channel to an output channel must traverse three switch modules 102, one switch module 102 in each of the three arrays 190-1, 190-2, 190-3. In a folded arrangement, a single switch module is traversed if the input channel and the output channel are connected to the same switch module. Otherwise, a path traverses three switch modules as in the case of the unfolded arrangement. In a double-folded arrangement, a path from an input channel to an output channel may traverse a single switch module, two switch modules or three switch modules, as described in U.S. patent application Ser. No. 10/223,222 filed on Aug. 20, 2002, and titled "Modular High-Capacity Switch".

FIG. 2 illustrates a (known) classical time-space-time switch 200, often referenced as T-S-T switch, which has been extensively used for time-division-multiplexing (TDM) switching with the space stage implemented as electronic switches. An incoming optical signal in a channel is converted to an electrical signal by a first optical-to-electric converter 212. The electrical signal is then received by a time-switching module 206 in a first switching array 290-1.

A time-switching module 206 receives signals that are arranged in a time frame having a predefined number of time slots. A signal contained within a time-slot has a predefined destination.

The space switch 204 may be electronic or optical. When space switch 204 is implemented as an optical switch, the electrical signal at the output of the time-switching module 206 is converted to an optical signal by a first electrical-to-optical converter 214. The resultant optical signal is then received by an optical space switch 204. After switching in the optical space switch 204, the switched optical signal is converted to an electrical signal by a second optical-to-electrical converter 222 and subsequently received by a time-switching module 206 in a second switching array 290-2. The electrical signal at the output of the time-switching module 206 is converted to an optical signal by a second electrical-to-optical converter 214. A scheduling processor (not illustrated) performs a time-slot matching process between the first switching array 290-1 and the second switching array 290-2.

An adapted version of the classical time-space-time switch 200 of FIG. 2 used for packet switching is also well known in the art (see, for example U.S. Pat. No. 5,168,492 issued on Dec. 1, 1992, to Beshai et al., U.S. Pat. No. 5,475,679 issued on Dec. 12, 1995, to Munter, and U.S. Pat. No. 5,745,486 issued on Apr. 28, 1998, to Beshai et al.)

FIG. 3 illustrates a known switch 300, having an architecture known as the space-time-space (S-T-S), which has been employed for TDM switching. Historically, the S-T-S TDM switch 300 of FIG. 3 has been less popular than its T-S-T counterpart (see FIG. 2) because a time-switching module 206 is quite simple to construct and the T-S-T uses fewer space switching modules and more time-switching modules in comparison with an S-T-S switch of the same capacity.

The S-T-S TDM switch 300 of FIG. 3 comprises two optical space switches 304A, 304B, connected by a switching array 390 of time-switching modules 306. The first optical space switch 304A has n>1 inlet ports and m.gtoreq.n>1 outbound ports while the second optical space switch 304B has m inbound ports and n outlet ports. Each outbound port transmits data to one of the time-switching modules 306 and each inbound port receives data from one of the time-switching modules 306. When optical space switching is employed, the S-T-S TDM switch 300 of FIG. 3 may be less expensive than the T-S-T TDM switch 200 of FIG. 2 because only one optical-to-electrical conversion and only one electrical-to-optical conversion are required.

The scalability of the T-S-T switch 200 of FIG. 2 and the S-T-S switch 300 of FIG. 3 is limited by the scalability of the space switching stages. The scalability can be enhanced significantly by replacing the pure time-switching modules 206, 306 by time-space switching modules, where each time-space switching module has several input ports and several output ports. A good example of a time-space switching module is a known common-memory switch to be described below.

A common-memory switch relies on massive data parallelism to enable high-speed data storage and retrieval. Data is stored in a common-memory comprising parallel memory devices which are identically addressed. The common-memory switch may have several input ports and several output ports. At any instant of time, only one input port may have exclusive write access to the common-memory, or only one output port may have a read access to the common-memory. In a common-memory switch, there is no internal congestion and input data is guaranteed a path to its desired output port. In one implementation, a time-frame having a predefined number of time slots is used to coordinate memory access among the input ports and the output ports.

Known common-memory switching devices use fixed size data blocks, such as ATM (asynchronous transfer mode) cells or STM (synchronous transfer mode) data blocks. For example, U.S. Pat. No. 5,144,619 titled "Common memory switch for routing data signals comprising ATM and STM cells", issued to Munter on Sep. 1,.sup.st 1992, describes a common memory switch that handles data segments of a fixed size. U.S. Pat. No. 6,118,792 titled "Method and Apparatus for a Flexible-Access Rate Common-Memory Packet Switch", issued on Sep. 12, 2000 to Beshai, describes a common-memory switch having a plurality of input ports and a plurality of output ports where the sum of the capacities of the input ports may exceed the internal capacity of the switch as determined by the speed of the common memory and the sum of the capacities of the output ports may also exceed the internal capacity of the switch. An implicit concentration stage is realized by adaptively allocating permissible access rates for each input port. Each input port transfers data segments of equal size to the common memory at specified time slots and the allocated access rate of each port is based on the fixed data-segment size. The allocated access rate for an input port applies to the total traffic received at the input port.

FIG. 4 illustrates a structure of an alternative S-T-S switch 400 adapted from the S-T-S switch 300 of FIG. 3. Rather than the single optical space switch 304 at the input stage, an input array 490-1 of optical space switches 304, each identical to the optical space switches 304A, 304B of FIG. 3, is used as an input stage. A middle stage is made up of a middle array 490-2 of 8.times.8 time-space switching modules 406, each implemented, for example, as a common-memory switch as described above. An output stage mimics the input stage by using an output array 490-3 of optical space switches 304. Although only two optical space switches 304 in each of the input array 490-1 and the output array 490-3 are shown to connect to the time-space switches of the middle stage, this arrangement is merely for simplicity of illustration. It should be understood that each of the optical space switches 304 in the input stage connects to all of the 8.times.8 time-space switching modules 406 in the middle stage and that each of the optical space switches 304 in the output stage connects to all of the 8.times.8 time-space switching modules 406 in the middle stage.

Time-space switches of a dimension larger than 8.times.8 may also be used. For example, the S-T-S switch 400 of FIG. 4 comprises eight input space switches 304, eight output space switches 304 and four 16.times.16 time-space switches 406. The S-T-S switch 400 can be viewed as a superposition of eight S-T-S switches 300 which interconnect through the four 16.times.16 time-space switches 406. Each input space switch 304 connects to each time-space switch 406, and each time-space switch 406 connects to each output optical switch 304. An input space switch 304 may connect to a time-space switch 406 through two wavelength channels and a time-space switch. 406 may connect to each output space switch 406-B through two wavelength channels. Alternatively, eight time-space switches 406 each of dimension 8.times.8 may be used to interconnect the input space switches 304 to the output space switches 304. An input space switch 304 then connects to a time-space switch 406 through one wavelength channel and a time-space switch 406 connects to each output space switch 406-B through one wavelength channel.

A time-space switching module operates under control of a switching schedule. Such a switching schedule may be defined for a time frame that includes a series of time slots. The switching schedule determines input-output connectivity during each time slot. Through such reconfiguration, a particular input channel may be connected to a given output channel during one time slot and connected to another output channel during another time slot.

FIG. 5 illustrates a structure of a combination switching node 500 that combines an S-T-S switch and an S-S-S switch in one structure. An input array 590-1 includes a number of optical space switches 504 each connected to all of the switches of a middle array 590-2. The middle array 590-2 includes channel switches 508 and baseband switches 510. The switches of the middle array 590-2 are, in turn, connected to all of the switches of an output array 590-3. Like the input array 590-1, the output array 590-3 includes a number of optical space switches 504. The combination switching node 500 of FIG. 5 thus allows switching of an entire channel from an inlet port to an outlet port through the channel switches 508 of the middle array 590-2, or switching of time-slotted signals, through the time-space switches (baseband switches 510) of the middle array 590-2.

Given that each channel accommodates an optical carrier of a particular wavelength modulated by an information signal, a distinction can be drawn between the channel switches 508 and the baseband switches 510 of the middle array 590-2 as follows. A given channel switch 508 receives an optical signal in a channel of a particular wavelength band from one of the space switches 504 in the input array 590-1 and transmits the entire optical signal carried by the channel to a configured destination one of the space switches 504 in the output array 590-3. The transmitted optical signal may be shifted to another wavelength band. A given baseband switch 510 receives a time-divided channel in a particular wavelength band from one of the space switches 504 in the input array 590-1 and transmits, for the duration of each time slot, the information signal, in a wavelength band appropriate to the configured destination one of the space switches 504 in the output array 590-3.

It is to be noted that, in the structure of FIG. 3 or FIG. 4, the space switches 304 may be electronic space switches, optical space switches, or a mixture of electronic and optical space switches. The channel switches 508 of the structure of FIG. 5 may also be electronic, optical, or a mixture of electronic and optical space switches.

FIG. 6 illustrates a switch structure 600 derived by rearranging the outer space switches and the channel switches (space switches) of the middle array 590-2 in structure of FIG. 5. In the. structure 600 of FIG. 6, a single, integrated, space-switch module 604 performs the functions of a space switch 504 in the input array 590-1, a channel switch 508 in the middle array 590-2 and a space switch 504 in the output array 590-3 of the combination switching node 500 of FIG. 5. The integrated modules 604 are connected in a mesh structure. In FIG. 7, connections of the switch structure 600 are shown between the integrated modules 604 and a baseband switch modules 610A. Another baseband switch module 610B is likewise connected. For simplicity of illustration, the baseband switch modules 610 and related connections are omitted in FIG. 6. Each baseband switch module 610 is connected to each of the integrated modules and functions as one of the baseband switches 510 of the middle array 590-2, in a folded structure.

A channel or time-shared connection from an input of a first integrated module 604 to an output of a second integrated module 604 may be switched internally, if the second integrated module 604 is also the first integrated module 604. A channel connection may either traverse only the first and second integrated modules 604, or be routed through an intermediate integrated module 604. A time-shared connection from an input of an integrated module 604 to an output of another integrated module 604 can be switched through one of the baseband modules 610.

A network 800 of bufferless switches 804A, 804B, 804C, 804D (individually or collectively 804) and subtending edge nodes 805P, 805Q, 805X, 805Y (individually or collectively 805) is illustrated in FIG. 8. In order for a time slot within a time frame to be switched, at a bufferless switch 804, to the appropriate sink node, timing is critical. The source node or switch of the time frame should be "time-locked" to the bufferless switch 804 that is to perform the switching. The time-locking necessary to properly switch a time slot within a time frame across a path including more than one bufferless switch 804 can be difficult to establish, as will be illustrated with reference to FIG. 9.

To establish time-locking between a first network element (node or switch) and a second network element, one of the network elements may be designated a master and the other network element may be designated a slave. Each of the two network elements includes a time counter. Each time counter has a predefined word-length; 24 bits for example. Such time counters are cyclic, each counting up to a maximum value before resetting to zero. Alternatively, such counters , called down counters, could start at the maximum value and count down to zero (often called down counters). The slave time counter and the master time counter may have the same number of bits and may be controlled by a clock running at the same rate. Through control signaling, the slave network element can time-lock the slave time counter to the master time counter.

The master time counter may, for instance, define a master cycle with a starting point of time zero. Consider a scenario wherein the slave time counter is locked to the master time counter. If the slave network element starts to send a time frame at time zero on the slave time counter, the time frame will, because of a distance traveled by the time frame, arrive at the master network element some time after time zero on the slave time counter. Once the time-locking procedure has been applied, the time at which the time frame is transmitted by the slave network element is adjusted such that the time frame sent from the slave network element arrives precisely at a designated time with respect to the master time counter, as will be detailed below. This requires that the cycle duration of each time counter exceed the round-trip delay between the slave and master time counters.

In order to effect time-locking, the slave network element may send a control signal to the master network element, where the control signal indicates the reading of the slave time counter at the time of sending. The master network element, upon receiving the control signal, may compare the indication of the reading on the slave time counter with the reading on the master time counter at the time of receipt of the control signal. The master network element may then send the slave network element a control signal indicating an amount by which to adjust the slave time counter in order that a time frame sent at a given slave time counter reading will arrive at the master network element at a time at which the reading of the master time counter is identical to the given slave time counter reading. In one implementation, described in the aforementioned U.S. patent application Ser. No. 09/286,431, the master network element sends the master time counter reading, which is used by the slave controller to reset the slave time counter.

FIG. 9 illustrates an arrangement of time counters wherein one of the bufferless space switches 804C and one of the edge nodes 805P of FIG. 8 (acting as slave network elements) are independently time-locked to another bufferless space switch 804A (acting as a master network element).

Returning to FIG. 8, it is notable that time counters at the bufferless space switches 804C and 804D may time-lock to a time counter at the bufferless space switch 804A, time counters at the nodes 805P, 805Q can time-lock to time counters at the bufferless space switches 804C and 804A and time counters at the edge nodes 805X, 805Y can time-lock to time counters at the bufferless space switches 804D, 804A. Further, while time-locked to a time counter at the bufferless space switch 804A, time counters at the bufferless space switches 804C, 804D and edge nodes 805P, 805Q, 805X, 805Y cannot time-lock to a time counter at the bufferless space switch 804B, except by coincidence.

Let bufferless space switch 804A be a master network element. Then time counters at bufferless space switches 804C and 804D may time-lock directly to a time counter at bufferless space switch 804A. Time counters at edge nodes 805P, 805Q can time-lock directly to a time counter at bufferless space switch 804C, and hence be time-locked to the time counter at bufferless space switch 804A. Time counters at edge nodes 805X, 805Y can time-lock directly to the time counter at bufferless space switch 804D, and hence be time-locked to the time counter at bufferless space switch 804A. In this case, the time counters at bufferless space switches 804C and 804D, would not time-lock to a time counter at bufferless space switch 804B except by coincidence. It follows that, while time-locked to a time counter at the bufferless space switch 804A, the time counters at edge nodes 805P, 805Q, 805X, 805Y would not time-lock to the time counter at bufferless space switch