WO2002007348A1

WO2002007348A1 - Hybrid optical shared protection ring

Info

Publication number: WO2002007348A1
Application number: PCT/US2001/022298
Authority: WO
Inventors: Ming-Jun Li; Mark J Soulliere
Original assignee: Corning Incorporated
Priority date: 2000-07-17
Filing date: 2001-07-16
Publication date: 2002-01-24
Also published as: AU2001276929A1

Abstract

A wavelength add drop multiplexer (WADM) has a unique 4X4 switching matrix configuration that is capable of supporting a 2-fiber ring architecture. A first 4X4 switching matrix is optically coupled to a first incoming fiber carrying a first incoming optical transport signal of n channels (OTSn) and a second incoming fiber carrying a second incoming OTSn. The first 4X4 switching matrix routes a working channel of the first incoming OTSn and a protection channel of the second incoming OTSn based on a desired switching configuration. The WADM further includes a second 4X4 switching matrix optically coupled to the first incoming fiber and the second incoming fiber. The second 4X4 switching matrix routes a protection channel of the first incoming OTSn and a working channel of the second incoming OTSn based on the desired switching configuration. The WADM also has a control system for establishing the switching configuration of the matrices. Routing working traffic and protection traffic from a single fiber to a pair of 4X4 switching matrices ultimately allows for simultaneous protection from multi-channel failures and single channel failures.

Description

HYBRID OPTICAL SHARED PROTECTION RING

Background Of The Invention

1. Field of the Invention

The present invention generally relates to optical communication systems, and more particularly to a wavelength add drop multiplexer that enables a hybrid optical shared protection ring to operate outside of conventional SONET and SDH protection schemes.

2. Technical Background

In the rapid development of optical communication systems, networking architectures have become increasingly complex. Ring topologies have arisen to provide a number of networking elements with the ability to both listen and transmit on optical channels within the optical ring. In basic ring topologies, consecutive nodes are connected by point-to-point links which are arranged to form a single closed path or ring. Information is transmitted from node to node around the ring, and the interface at each node is an active device that has the ability to create and accept messages. The interface is part of a wavelength add drop multiplexer (WADM) that routes signals to and from the client network elements. Thus, the WADM not only provides the client network elements a user attachment point, but also enables the node to function as an active repeater for re-transmitting messages that are addressed to other nodes.

A number of implementation considerations must be taken into account when configuring a ring network. First, rings must be physically arranged so that the nodes provide communication between the desired client network elements. If fact, certain client network elements may desire communication with one another while others may not. Optical signals therefore typically carry numerous channels to provide users with the desired communication links. Furthermore, the number of fibers being used to transfer the signals from node to node drastically effects the capacity of the ring and the structure of the WADM. For example, rings having a four-fiber architecture typically will transport working traffic on two of the fibers, and protection traffic on the other two fibers. The working traffic fibers will carry traffic in opposing directions, as will the protection traffic fibers. While this type of four-fiber architecture is acceptable for long distance networks, it is desirable to reduce the number of fibers required for "metro" networks. Such a reduction would ultimately provide more efficient bandwidth usage.

Another consideration to be taken into account is the possibility of network failures. The American National Standards Institute (ANSI) has released a collection of standards for a synchronous optical network (SONET) that addresses the growing problem of single channel failures. Similar standards have been developed for

Synchronous Digital Hierarchy (SDH) network protection architectures. These standards provide signaling protocols for various types of single channel optical networks but fail to address the WADM with any specificity. This is crucial since, as already discussed, the WADM is coupled to one or more single channel client network elements and provides access to the ring. Another problem with structuring an optical ring around SONET or SDH standards, is the possibility of transmitting data which is not SONET or SDH based. For example, Gigabit Ethernet signals transmitted to digital clients often do not fall within established network protection standards. Thus, it is desirable to provide a WADM and an optical shared protection ring capable of functioning within or out of standardized protocols.

As already discussed, it is well known that for an optical transmission section signal of n channels (OTSn) being transferred between two or more client network elements, a number of component, equipment, and device problems can cause either multi-channel or single channel failures. For example, multi-channel failures occur when either a fiber cut or a cable cut causes a loss in more than one optical channel of the OTSn. Conventional SONET and SDH approaches have involved the construction of the ring such that protection capacity is switched in to restore working traffic. This is commonly termed "shared protection" because the same wavelengths are used for different client links. A single channel failure typically occurs as a result of a WADM switching matrix failure or a client network element transmitter/receiver failure. To avoid these single channel failures, redundant transmitters and receivers are typically used at the client network element site. This is commonly termed 1+1 linear protection switching and has been described for both SONET and SDH architectures. The difficulty with conventional approaches, however, is that outside of SONET/SDH domain, there has been no way to protect for both multi-channel and single channel failures at the same time. In fact, due to the lack of available guidance in the field of WADM structures, protection has been challenging even within the SONET/SDH domain. This is unfortunate because a hybrid optical shared protection ring would simultaneously support a plurality of client service protection options. It is. therefore desirable to provide protection against all optical single points of failure. It is also desirable to provide simultaneous per channel protection options. This would allow many levels of client traffic protection to be configured, and these configurations could exist on the same ring on a per channel basis.

Summary Of The Invention

The present invention is a wavelength add drop multiplexer having a 4X4 switching matrix configuration, and is capable of supporting a 2-fiber ring architecture.

A first 4X4 switching matrix is optically coupled to a first incoming fiber carrying a first incoming optical transmission section signal of n channels (OTSn) and a second incoming fiber carrying a second incoming OTSn. The first 4X4 switching matrix routes a working channel of the first incoming OTSn and a protection channel of the second incoming OTSn based on a desired switching configuration. The multiplexer further includes a second 4X4 switching matrix optically coupled to the first incoming fiber and the second incoming fiber. The second 4X4 switching matrix routes a protection channel of the first incoming OTSn and a working channel of the second incoming OTSn based on the desired switching configuration. The wavelength add drop multiplexer supports multiple pairs of 4Xx4 switching matrices to provide protection for multiple optical channels. The wavelength add drop multiplexer also has a control system for establishing the switching configuration of the switching matrices. Routing working traffic and protection traffic from a single fiber to a pair of 4X4 switching matrices ultimately allows for simultaneous protection from multi-channel failures and single channel failures.

Further in accordance with the present invention, an optical shared protection ring for interconnecting client network elements is provided. Specifically, the ring includes a plurality of wavelength add drop multiplexers optically coupled to the client network elements, and a two-fiber architecture optically coupled to the wavelength add drop multiplexers. The two-fiber architecture transports a set of working channels and protection channels of a first OTSn between the wavelength add drop multiplexers in a first direction. The two-fiber architecture further transports a set of working channels and protection channels of a second OTSn between the wavelength add drop multiplexers in a second direction. The result is a hybrid optical shared protection ring that provides simultaneous protection from. multi-channel failures and single channel failures regardless of whether or not the transported signals are SONET/SDH based. It is to be understood that both the foregoing general description and the following detailed description are merely exemplary of the invention, and are intended to provide an overview or framework for understanding the nature and character of the invention as it is claimed. The accompanying drawings are included to provide a further understanding of the invention, and are incorporated in and constitute part of this specification. The drawings illustrate various features and embodiments of the invention, and together with the description serve to explain the principles and operation of the invention.

Brief Description Of The Drawings

The various advantages of the present invention will become apparent to one skilled in the art by reading the following specification and appended claims, and by referencing the following drawings, in which:

FIG. 1 is a schematic illustration of an optical shared protection ring implementing the presently preferred wavelength add drop multiplexer in accordance with the principles of the present invention;

FIG. 2 is a diagram of a wavelength add drop multiplexer in accordance with the principles of the present invention; FIG. 3 is a diagram of a wavelength add drop multiplexer configured for one hundred percent add drop connectivity in accordance with the principles of the present invention.

FIG. 4 is a diagram of a wavelength add drop multiplexer configured as a through node in accordance with the principles of the present invention;

FIG. 5 is a diagram of a wavelength add drop multiplexer showing n channels in accordance with the principles of the present invention;

FIG. 6 is a block diagram showing the 4X4 nature of the wavelength add drop multiplexer switch matrices in accordance with the principles of the present invention; FIG. 7 is a schematic illustration of an optical shared protection ring with a single fiber failure switching in accordance with the principles of the present invention; FIG. 8 is a diagram of Node B shown in FIG. 7;

FIG. 9 is a block diagram of a switch matrix controller in accordance with the principles of the present invention; FIG. 10 is an illustration of a node signaling system for the failure and repair of a single fiber failure as shown in FIG. 7 in accordance with the principles of the present invention;

FIG. 11 is a schematic illustration of an optical shared protection ring with a cable failure undergoing adjacent node (or end-node) switching in accordance with the principles of the present invention;

FIG. 12 is a diagram of Node B shown in FIG. 11; FIG. 13 is a diagram of Node D shown in FIG. 11;

FIG. 14 is a schematic illustration of an optical shared protection ring with a cable failure undergoing non-adjacent node (or end-node) switching in accordance with the principles of the present invention;

FIG. 15 is a diagram of Node B shown in FIG. 14; FIG. 16 is a diagram of Node D shown in FIG. 14;

FIG. 17 is a diagram of a wavelength add drop multiplexer with a switch matrix failure wherein the multiplexer is configured for one hundred percent add drop capability; FIG. 18 is a diagram of a wavelength add drop multiplexer with a switch matrix failure wherein the multiplexer is configured as a through node in accordance with the principles of the present invention;

FIG. 19 is a diagram of a wavelength add drop multiplexer with a client network element transmitter failure wherein the multiplexer is configured in accordance with the principles of the present invention;

FIG. 20 is a diagram of a wavelength add drop multiplexer with a client network element receiver failure wherein the multiplexer is configured in accordance with the principles of the present invention; FIG. 21 is a diagram of a wavelength add drop multiplexer connected to provide three distinct levels of protection;

FIG. 22 is a diagram of a wavelength add drop multiplexer connected to a SONET bidirectional line switched ring client network element;

FIG. 23 is a schematic illustration of a first alternative optical shared protection ring with a fiber cut wherein the multiplexer is configured in accordance with the principles of the present invention;

FIG. 24 is a diagram of Node B shown in FIG. 23;

FIG. 25 is a schematic illustration of a second alternative optical shared protection ring with a cable cut wherein the multiplexer is configured in accordance ι with the principles of the present invention;

FIG. 26 is a diagram of Node B shown in FIG. 25;

FIG. 27 is a diagram of a pair of wavelength add drop multiplexers configured to interconnect two rings in accordance with the principles of the present invention;

FIG. 28 is a diagram of the wavelength add drop multiplexers shown in FIG. 27 wherein the multiplexers are configured for one hundred percent interconnection capability;

FIG. 29 is a diagram of the wavelength add drop multiplexers shown in FIG. 27 wherein the multiplexers are configured to separate traffic traveling on the two rings; and FIG. 30 is a diagram of the wavelength add drop multiplexers shown in FIG. 28 wherein a working link cable cut has occurred. Detailed Description Of The Preferred Embodiments

Reference will now be made in detail to the present preferred embodiments of the invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same reference numerals will be used throughout the drawings to refer to the same or like parts.

Referring to FIG. 1, a hybrid optical shared protection ring in accordance with the present invention is shown generally at 40. It can be seen that the ring 40 interconnects client network elements 50, 52, 54, 56, and provides protection from all single points of failure. Node A includes client network element (NE) 50, client NE 56 and wavelength add drop multiplexer (WADM) 60. Node B includes client NE 52, client NE 54 and WADM 62. Specifically, the ring 40 is configured to provide a first communication link between client NE 54 and client NE 56. A second communication link exists between client NE 50 and client NE 52. The ring 40 has a plurality of WADMs 60, 62, 64, 66 optically coupled to the client network elements. The ring 40 also has a two-fiber architecture 42 optically coupled to the WADMs 60, 62, 64, 66. As will be described in greater detail below, the two-fiber architecture 42 transports a working channel and a protection channel of a first optical transmission section signal of n channels (OTSn) between the WADMs in a first direction (shown as counter- clockwise). The traffic traveling in the first direction is carried on fiber 1. The two- fiber architecture 42 further transports a working channel and a protection channel of a second OTSn between the WADMs in a second direction (shown as clockwise). The traffic traveling in the second direction is carried on fiber 2.

Transporting working and protection traffic on the same fiber allows conventional four-fiber architectures to be reduced to two fibers. The cost savings is significant and the resulting ring 40 is more suitable for metro networks. As will be discussed below, the ring 40 provides client network elements with protection from multi-channel failures as well as single channel failures with an extremely efficient design. It is important to note that the structure of the WADMs 60, 62, 64, 66 is also critical to the ability to operate under a two-fiber architecture.

Turning now to FIG. 2, a portion of the preferred WADM 70 is shown. As already discussed, the OTSn signals traveling on either fiber can transport up to n optical channels on n different wavelengths. Half of these wavelengths (n/2, presuming that n is even) are reserved for working traffic, and the other half are reserved for protection traffic. While many wavelength assignment schemes are available, the preferred scheme is selected such that the wavelengths of working traffic traveling clockwise around the ring 40 must be disjoint from the wavelengths of the working traffic traveling counterclockwise around the ring 40. Thus, any working wavelength λ_j traveling counterclockwise on fiber 1 has,

l ≤ j ≤ l .

2

Any protection wavelength λ_k also traveling counterclockwise on fiber 1 has,

-+l ≤ k ≤ n .

2 Any protection wavelength λ_j traveling clockwise on fiber 2 has,

l ≤ j ≤ - .

2

Any working wavelength λ also traveling clockwise on fiber 2 has,

- + l ≤ k ≤ n .

2

It is also presumed that a fixed relation holds between λ_j and λ_k:

k = j +- .

2

If a specific working wavelength λ_j is chosen between two client network elements (e.g. from client NE 50 to client NE 52), then the associated working wavelength in the return direction (e.g. from client NE 52 to client NE 50) is λ_k. This wavelength assignment scheme eliminates the need for wavelength converters. FIG. 1 and others below use two of the many possible optical channel connections to illustrate the self-healing functionality of the present invention. Client NE 50 has a bi-directional working optical channel that reaches client NE 52 via Nodes A and B. Live traffic (i.e. "working" traffic) is placed on wavelength λ_j between client NE 50 and client NE 52. A replica of the live traffic (i.e. "protection" traffic) is simultaneously on wavelength λ_k between client NE 50 and client NE 52. As already discussed, both the working and protection traffic travel from Node A to Node B on the same fiber. Similarly, the return traffic from client NE 52 to client NE 50 has a working optical channel on wavelength λ_k and a protection optical channel on wavelength λ_j.

Another bi-directional optical channel is illustrated traveling from client NE 54 to client NE 56 via Nodes B, C, D and A. In the illustrated example, λ_j is reused for working traffic on the counterclockwise spans, and λ is reused for working traffic on the clockwise spans.

Returning to FIG. 2, it can be seen that the WADM 70 has two multi-channel ports 72, 74, with each port having n incoming and outgoing OTSn. As best seen in FIG. 5, each WADM 70 has n 4X4 single channel non-blocking optical switch matrices 71, 73 (one for each of the n groups of optical channels). It is important to note that all of the equipment typically present, such as optical amplifiers, are not shown. Nevertheless, the functions necessary to illustrate the principles of operation are given. The multi-channel OTSn signal arriving at the WADM 70 is de-multiplexed, and each of the n optical channels are routed to a single wavelength switch matrix. An optical channel arriving from port 72 may either travel through towards port 74, or drop, have its optical channel overhead terminated, and have its constituent client signal travel towards a client NE. For a two-fiber WADM with one hundred percent add drop capability, one could have up to 2n separate pairs of client signals.

Thus, each WADM 70 includes a first 4X4 switching matrix 71 optically coupled to a first incoming fiber 75 carrying a first incoming OTSn and a second incoming fiber 76 carrying a second incoming OTSn. The first 4X4 switching matrix 71, routes a working channel of the first incoming OTSn and a protection channel of the second incoming OTSn based on a desired switching configuration. Similarly, a second 4X4 switching matrix 73, is optically coupled to the first incoming fiber 75 and the second incoming fiber 76. The second 4X4 switching matrix 73 routes a protection channel of the first incoming OTSn and a working channel of the second OTSn based on the desired switching configuration. The WADM 70 further includes a control system 90 for establishing the switching configuration of the matrices 71, 73. As best shown by FIG. 6, the switch matrices 71, 73 are in fact 4X4. Turning now to FIG. 3, a WADM 62 configured for one hundred percent add drop capability is shown under normal (i.e. failure free) conditions. It is important to note that for a multichannel OTSn signal carrying n optical channels, there can be as many as n/24X4 matrices 71 and also n/2 more 4X4 matrices 73. For simplicity, only two are shown in FIG. 3. The illustrated multiplexer corresponds to Node B from FIG. 1. Thus, the control system 90 has configured the WADM 62 to function as part of an add drop node. Specifically, Node B has one hundred percent add drop capability at an optical channel using wavelength λ_j eastbound and λ_k westbound, hi other words, the first 4X4 switching matrix 71 drops the working channel of the first incoming OTSn to a working output and drops the protection channel of the second incoming OTSn to a protection output. It can further be seen that the first 4X4 switching matrix 71 is optically coupled to a first outgoing fiber 77 and a second outgoing fiber 78. The first 4X4 switching matrix 71 adds a protection input to the first outgoing fiber 77 and adds a working input to the second outgoing fiber 78. The result is that the first outgoing fiber 77 carries a protection channel of a first outgoing OTSn and the second outgoing fiber 78 carries a working channel of a second outgoing OTSn.

The second 4X4 switching matrix 73 drops the protection channel of the first incoming OTSn to a protection output and drops the working channel of the second incoming OTSn to a working output. It will be appreciated that the second 4X4 switching matrix 73 is further optically coupled to the first outgoing fiber 77 and the second outgoing fiber 78. Under normal conditions, the second 4X4 switching matrix 73 adds a working input to the first outgoing fiber 77 and adds a protection input to the second outgoing fiber 78. The result is that the first outgoing fiber 77 carries a working channel of a first outgoing OTSn and the second outgoing fiber 78 carries a protection channel of a second outgoing OTSn.

It is important to note that at other working wavelengths, Node B may have just one add drop (fifty percent), or no add drops (a through node). In other words, the add drop configuration of any WADM node may differ from wavelength to wavelength. In

FIG. 3, the signals chosen by the client network elements are circled. In this configuration, the protection add drop traffic from client NEs 52 and 54 are placed onto the protection capacity of the WADM 62. From the WADM's perspective, the client NE 52 and client NE 54 protection traffic is "extra traffic"~namely, traffic that gets transported, but is discontinued when the optical shared protection ring performs self- healing. This implies a ranking of failures for those times when simultaneous failures exist. A multi-channel failure requiring a ring switch will take precedence over a single channel failure requiring the use of the extra traffic connection. This makes sense, because protecting a multi-channel failure heals more traffic than protecting one (or even several) single channel failures.

It is important to note that the WADM can also be configured to function as part of a through node, as shown in FIG. 4. This is WADM 66 as shown in FIG. 1. Here, it can be seen that the first 4X4 switching matrix 71 routes the working channel of the first incoming OTSn directly to the second outgoing fiber 78 and routes the protection channel of the second incoming OTSn directly to the first outgoing fiber 77.

While the present invention has been described above as operating under normal conditions, it will be appreciated that protection is provided from multi-channel failures and single channel failures. Generally, a multi-channel failure can occur as a result of a single fiber cut as shown in FIG. 7, or a cable cut to be discussed below. Specifically, it can be seen that a failure of fiber 1 between nodes A and B has occurred. Client NE 52 has lost its working optical channel from client NE 50. It has also lost the corresponding protection optical channel, since both were on the same fiber. This is a multi-channel failure, so the nodes implement optical shared protection. This may be optical multiplex section shared protection, where multichannel failures are detected and signaled, but switching is done on a per channel basis. This also may be optical channel shared protection, whereby single channel failures are detected (even when caused by multichannel failures) and signaled, and switching is done on a per channel basis.

WADM 62 places the working add signal from client NE 52 onto protection capacity traveling away from the failure, namely through Nodes C and D to Node A. WADM 60 performs a similar action, such that the interrupted traffic is restored. This rearrangement occurs for each set of optical channels at other wavelengths that were interrupted by the fiber failure. Working traffic not affected by the fiber failure (e.g. client NE 56 to client NE 54 is not interrupted). The protection copies of the traffic that existed as extra traffic, however, have been disconnected.

It will be appreciated that the above self-healing requires coordinated action among Nodes A, B, C and D. Signaling among these nodes coordinates the action.

One approach to the signaling uses messages sent via an optical supervisory channel that is terminated at every node. A signaling system heavily based on SDH shared protection rings is shown in FIG. 10 at 80. This illustrates optical multiplex section shared protection. Nodes A through D are shown in Figure 10, with a timeline traveling from top to bottom. Each node sends a message that includes: the current request for another node to switch; the source and destination nodes for the request; whether the request is being sent across the span having the problem (the "short" path), or the long way around the ring (the "long" path); and the switching state of the source node.

At time To, the ring is failure-free, and there are no switch requests. At time Ti, Node B detects a failure of the incoming multichannel signal (termed an Optical Multiplex Section signal within ITU-T Recommendation G.872). Node B disconnects its extra traffic, and sends a "signal fail - ring switching" request to Node A on both the long and short paths around the ring. Node C receives a switch request message from Node B. Since neither the source nor destination addresses in the message match Node C, it knows that it is an "intermediate node." As an intermediate node in this example, Node C (1) disconnects its extra traffic, (2) configures its protection wavelengths to be through channels, and (3) bidirectionally passes through all received switch request messages. Node B's request then continues to Node D, which takes the same intermediate node actions.

At time T_1? Node A is not aware that there is a problem. After receiving a switch request message from Node B, it disconnects its extra traffic, performs a headend switch, and sends a switch request back to Node B. When Node B receives this request, it connects its affected add signals to outgoing protection capacity (head end switch), and connects its affected drop signals to incoming protection capacity (tail end switch). The state of Node B at that point is shown in Figure 8. After receiving the head and tail end switched indication from Node B, Node A performs a tail-end switch. The signal between client NEs 50 and 52 is thus restored. Optical switch matrices that can complete their switching in the millisecond range, combined with fast signaling processing, should allow ring protection at comparable speeds to SONET/SDH equipment. This architecture differs from conventional SONET/SDH shared protection rings in that the head-end signal (the signal that is added to the ring) for SONET/SDH is "bridged," i.e. split and sent on both working capacity towards the failure and protection capacity away from the failure. In a conventional optical shared protection ring, the signal is not optically split to create a bridge, because a 50/50 split introduces a 3 dB optical penalty onto the split signal. The entire optical signal is instead switched to protection capacity traveling away from the failure. At time T₂, the fiber failure clears. The ring does not immediately revert back to its normal state, but waits for a preset "wait-to-restore" time. This prevents the ring from reconfiguring during a repair period where valid-looking signals may briefly appear at the nodes adjacent to the failure. At time T₃ the wait-to-restore time expires, and the ring switches are taken down in an order that prevents misconnected traffic. The extra traffic needs to be restored as well. In order to assure that the extra traffic is not momentarily misconnected during the final signaling sequence, an explicit message allowing extra traffic to return may be necessary. Such a message is presumed at time T4. Alternatively, timers are set after receiving "No Request" messages from adjacent nodes, and extra traffic is reconnected after the timers expire. The example of FIG. 10 shows that certain software processing is necessary to control the state of the switch matrices. FIG. 9 provides a conceptual view of the switch matrix control system 90 needed at each WADM. It can be seen that the control system 90 takes as input the detected incoming failures 91 from both ports, any locally detected equipment failures 92, any signaled messages from other Nodes 93, and network management information such as ring maps 94 and external user commands

95. From all these inputs, the switch matrix control system 90 uses its internal ring algorithm to send out switch commands to the various matrices.

As already mentioned, multi-channel failures can also result from a cable cut. Shared protection can be effected under this scenario by either performing adjacent node switching or non-adjacent node switching. FIG. 11 demonstrates the adjacent node switching scenario. It can be seen that the cable cut causes failures of both fibers 1 and 2 at a location not adjacent to either of the add drop nodes. This example also presumes that WADM 66 and WADM 64 (i.e. the nodes adjacent to the cable cut) perform a loop back switch, similar to what is done in SONET and SDH based shared protection rings.

It will be appreciated that the cable cut affects the working optical channel traveling between client NE 54 and client NE 56. The self-healing occurs by looping the affected working optical channels away from the failure onto the protection capacity, and eventually placing those optical channels back onto working capacity. In this example, the optical channel from client NE 54 to client NE 56 takes a routing of Node B to Node C, back to Node B, then Node A, Node D, then back to Node A. Nodes not adjacent to the failure, such as Nodes A and B, must therefore through- connect their protection capacity. All connections designated as extra traffic are therefore disconnected. FIG. 12 demonstrates that the protection copy of traffic between client NE 50 and client NE 52 is disconnected (though the working traffic remains). Thus, it can be seen that the adjacent nodes take all the optical channels originally destined for the working OTSn towards the failure and place them on the protection OTSn away from the failure. The switching action for Node D in FIG. 11 is shown in FIG. 13.

As already noted, another response to a cable cut can be non-adjacent node (or end-node) switching as shown in FIG. 14. Once again, the cable cut between Nodes C and D has interrupted both the working and protection versions of the traffic between client NE 54 and client NE 56. When using adjacent node switching (FIG. 11), it will be appreciated that the optical channels between client NE 54 and client NE 56 pass twice through Nodes A and B. A "head-end" and "tail-end" switch with WADM 60 and WADM 62 simplifies the routing, as shown in FIG. 14. The switch matrix configuration for WADM 62 is shown in FIG. 15. It will be appreciated that non- adjacent node switching disconnects extra traffic. An advantage of non-adjacent node switching over adjacent node switching is that the longest restoration route can be no more than the number of ring spans minus one. This is the same as the longest possible working route, so no special engineering is needed for the longest restoration route. Furthermore, FIG. 16 demonstrates that through nodes (even those adjacent to the failure), are not obliged to do any loop back switching.

While the above discussion of shared protection has focused upon multi-channel failures, it will be appreciated that the present invention also provides protection from single channel failures that are not covered by a conventional shared protection ring. Single channel failures can result from failures such as switch matrix failures, client NE transmitter failures, and client NE receiver failures. FIG. 17 demonstrates operation of the invention in response to a failure of the λ_j optical switch matrix 71. It will be appreciated that failure of matrix 71 actually impacts two different optical channels. Specifically, the dropped working optical channel from port 1, as well as the added working channel going to port 2 are interrupted. It will be appreciated that a conventional optical shared protection ring (either optical multiplex section shared protection or optical channel shared protection ring) would not heal this failure. In the hybrid optical shared protection ring, no switching action is performed by the WADMs. Instead, the client NEs 52 and 54 that lose their working copies of their signals, switch electrically to their protection copies. The protection copies are transported as extra traffic on the protection capacity of the conventional shared protection ring.

For the configuration shown in FIG. 1, a switch matrix failure in Node B (as discussed above) causes electrical protection switching within client NEs 50 and 56. As shown in FIG. 18, a switch matrix failure occurring in a through node such as Node D causes electrical protection switching within client NE 56, but no optical switching (for an optical multiplex section shared protection ring). Note that a multichannel failure could simultaneously invoke optical shared protection along with client level electrical protection switching. To prevent this competition, the electrical protection switching could delay its response to allow time for optical protection switching to occur. FIG. 19 shows WADM 62 for a transmitter failure in client NE 52. It will be appreciated that this also causes a single channel failure that a conventional optical shared protection ring would not heal. By providing 1+1 linear protection switching at the client NEs, and by transporting the protection copy of the client traffic as extra traffic on the optical shared protection ring, this failure can be healed without affecting the optical shared protection ring. Thus, client NE 52 detects the failed signal and performs an electrical switch to the protection copy. For the configuration used in FIG. 1, a transmitter failure within client NE 52 causes electrical protection switching within client NE 50.

FIG. 20 shows WADM 62 in the presence of a receiver failure in client NE 52. Once again, this failure is not protected by a conventional optical shared protection ring. The hybrid ring architecture heals this failure by having the client NE 52 detect the failed signal and perform an electrical switch to the protection copy. For the configuration used in FIG. 1, a receiver failure within client NE 52 causes electrical protection switching within client NE 50.

It is important to note that the present invention enables simultaneous support to different client network elements that are each using different levels of protection. For example, the first level of protection includes transporting signals as working traffic with both conventional optical shared protection and 1+1 linear protection. This provides maximum protection from both multi-channel and single channel failures. FIG. 21 demonstrates a WADM 100 connected to support client NE 110 in such a manner. It will be appreciated that client NE 110 can therefore be fully protected for non-SONET or SDH signals such as Gigabit Ethernet signals.

A second level of protection can be provided by transporting signals as work traffic with only optical shared protection. This type of connection is provided to client NE 120 and protects against multi-channel failures such as fiber cuts, cable cuts, and optical multiplexer/de-multiplexer failures. It will further be appreciated that a third level of protection can be provided by transporting signals as extra traffic with no protection. Such a connection is provided to client NE 130. Such a connection may be desirable for SONET or SDH based clients.

Turning now to FIG. 22, a fourth level of protection is shown for client NE 140, such as a SONET 2-fiber bi-directional line switched ring node. It can be seen that the client NE 140 is connected to the working wavelength of a WADM 100. The client NE 140 will therefore provide its own protection. In order to avoid competing protection schemes, the WADM 100 would need to suspend any reconfigurations of the working traffic connections during a failure. This capability is analogous to the non-pre- emptable unprotected traffic (NUT) feature of SONET BLSRs. A disadvantage of this configuration is that a failure of an optical switch matrix leads to isolation of the client NE 140.

It will be appreciated that the WADM of the present invention provides for alternative shared protection approaches. Specifically, FIG. 23 shows a first alternative approach to healing a fiber cut. It can be seen that the hybrid optical shared protection ring 40' retains the working optical channel on the other fiber that was not affected (i.e., the traffic from client NE 50 to client NE 52). This traffic is restored the long way around the ring, but the working traffic from client NE 52 back to client NE 50 is retained. A close up of WADM 62 is shown in FIG. 24. This approach means that WADM 60 would not have to perform a tail-end switch. In fact, it is simpler for both nodes to perform head-end and tail-end switches, as in the preferred embodiment, regardless of whether the failure is unidirectional or bi-directional. There may also be operational benefits to having both the "go" and "retum" services travel over matching facilities regardless of whether conditions are normal or failed.

FIG. 25 shows a second alternative embodiment of the present invention as ring 40". Comparing this approach with the embodiment shown in FIG. 7, it can be seen that the working "go" and "return" channels are on the same wavelength (λ_j), instead of on different wavelengths. Similarly, the protection go and return channels are both on a wavelength (λ_j) different from the working wavelength. A close up of WADM 62 is shown in FIG. 26. Specifically, a fiber cut on port 1 of Node B results in the protection add drops of client NE 52 being connected onto the protection capacity of port 2. This is to be contrasted with the preferred embodiment wherein the same fiber cut results in the working add drops of client NE 52 to be connected on to the protection capacity of port 2.

It will be appreciated that the WADM in FIG. 26 would sometimes drop protection add drop connections (e.g. to client NE 54) while keeping other protection add drop connections (e.g. to client NE 52). The WADM 62, and hence the optical shared ring 40", would not operate as a conventional optical shared protection ring whereby extra traffic is disconnected during failures. This can be viewed as a disadvantage as compared to the preferred embodiment. A second distinction with respect to the second alternative embodiment is that client NE 52 must also throw its electrical switch. This could potentially add to the service restoration time.

Furthermore, a client NE desiring only optical shared protection cannot be easily supported. For example, the WADM 62 would have to convert the add drop wavelength λj to λ_k in order to use the ring protection capacity.

Turning now to FIG. 27, it will be appreciated that the present invention provides for the interconnection of two fiber rings. It can be seen that as in the individual ring architecture already described, the interconnecting WADMs use 4X4 switch matrices. For the interconnecting node of ring 1, the client NEs are replaced by the interconnecting node of ring 2, and vice versa. The working and protection link cables are preferably deployed through different routes to avoid being cut at the same time. FIG. 28 shows the connections inside the switch matrices 152, 154, 162, 164 for the case where traffic traveling on the two rings is interconnected. FIG. 29 shows the connections inside the switch matrices 152, 154, 162, 164 for the case where the traffic traveling on the two rings is separated. FIG. 30 shows the interconnection discussed above in the presence of a working link cable cut. In this failure situation, the traffics are connected through the protection link cable. It is important to note that no signing between the two rings is required. Those skilled in the art can now appreciate from the foregoing description that the broad teachings of the present invention can be implemented in a variety of forms. Therefore, while this invention has been described in connection with particular examples thereof, the true scope of the invention should not be so limited since other modifications will become apparent to the skilled practitioner upon a study of the drawings, specification, and following claims.

Claims

What is claimed is:

1. An optical shared protection ring for interconnecting client network elements, the ring comprising: a plurality of wavelength add drop multiplexers optically coupled to the client network elements; a two-fiber architecture optically coupled to the wavelength add drop multiplexers; and said two-fiber architecture transporting a working channel and a protection channel of a first optical transmission section transport signal of n channels (OTSn) between the wavelength add drop multiplexers in a first direction around the ring, the two-fiber architecture further transporting a wavelength channel and a protection channel of a second

OTSn between the wavelength add drop multiplexers in a second direction around the ring.

2. The optical shared protection ring of claim 1 wherein each wavelength add drop multiplexer (WADM) includes: a first 4X4 switching matrix optically coupled to a first incoming fiber carrying a first incoming OTSn and a second incoming fiber carrying a second incoming OTSn, the first 4X4 switching matrix routing a working channel of the first incoming OTSn and a protection channel of the second incoming OTSn based on a desired switching configuration; a second 4X4 switching matrix optically coupled to the first incoming fiber and the second incoming fiber, the second 4X4 switching matrix routing a protection channel of the first incoming OTSn and a working channel of the second incoming OTSn based on the desired switching configuration; and a control system for establishing the switching configuration of the matrices.

3. The optical shared protection ring of claim 2 wherein the control system configures the add drop multiplexer to function as part of an add drop node.

4. The ring of claim 3 wherein the first 4X4 switching matrix drops the working channel of the first incoming OTSn to a working output and drops the protection channel of the second incoming OTSn to a protection output.

5. The optical shared protection ring of claim 4 wherein the first 4X4 switching matrix is further optically coupled to a first outgoing fiber and a second outgoing fiber, the first 4X4 switching matrix adding a protection input to the first outgoing fiber and adding a working input to the second outgoing fiber such that the first outgoing fiber carries a protection channel of a first outgoing OTSn and the second outgoing fiber carries a working channel of a second outgoing OTSn.

6. The optical shared protection ring of claim 3 wherein the second 4X4 switching matrix drops the protection channel of the first incoming OTSn to a protection output and drops the working channel of the second incoming OTSn to a working output.

7. The optical shared protection ring of claim 6 wherein the second 4X4 switching matrix is further optically coupled to a first outgoing fiber and a second outgoing fiber, the second 4X4 switching matrix adding a working input to the first outgoing fiber and adding a protection input to the second outgoing fiber such that the first outgoing fiber carries a working channel of a first outgoing OTSn and the second outgoing fiber carries a protection channel of a second outgoing OTSn.

8. A wavelength add drop multiplexer for supporting a 2-fiber ring architecture comprising: a first 4X4 switching matrix optically coupled to a first incoming fiber carrying a first incoming optical transport signal of n channels (OTSn) and a second incoming fiber carrying a second incoming OTSn, the first 4X4 switching matrix routing a working channel of the first incoming OTSn and a protection channel of the second incoming OTSn based on a desired switching configuration; a second 4X4 switching matrix optically coupled to the first incoming fiber and the second incoming fiber, the second 4X4 switching matrix routing a protection channel of the first incoming OTSn and a working channel of the second incoming OTSn based on the desired switching configuration; and a control system for establishing the switching configuration of the matrices.

9. The add drop multiplexer of claim 8 wherein the control system configures the add drop multiplexer to function as part of an add drop node.

10. The add drop multiplexer of claim 9 wherein the first 4X4 switching matrix drops the working channel of the first incoming OTSn to a working output and drops the protection channel of the second incoming OTSn to a protection output.

11. The add drop multiplexer of claim 10 wherein the first 4X4 switching matrix is further optically coupled to a first outgoing fiber and a second outgoing fiber, the first 4X4 switching matrix adding a protection input to the first outgoing fiber and adding a working input to the second outgoing fiber such that the first outgoing fiber carries a protection channel of a first outgoing OTSn and the second outgoing fiber carries a working channel of a second outgoing OTSn.

12. The add drop multiplexer of claim 9 wherein the second 4X4 switching matrix drops the protection channel of the first incoming OTSn to a protection output and drops the working channel of the second incoming OTSn to a working output.

13. The add drop multiplexer of claim 12 wherein the second 4X4 switching matrix is further optically coupled to a first outgoing fiber and a second outgoing fiber, the second 4X4 switching matrix adding a working input to the first outgoing fiber and adding a protection input to the second outgoing fiber such that the first outgoing fiber carries a working channel of a first outgoing OTSn and the second outgoing fiber carries a protection channel of a second outgoing OTSn.

14. The add drop multiplexer of claim 8 wherein the controller configures the add drop multiplexer to function as part of a through node.

15. The add drop multiplexer of claim 14 wherein the first 4X4 switching matrix is further optically coupled to a first outgoing fiber and a second outgoing fiber, the first 4X4 switching matrix routing the working channel of the first incoming OTSn to the second outgoing fiber and routing the protection channel of the second incoming OTSn to the first outgoing fiber.

16. The add drop multiplexer of claim 14 wherein the second 4X4 switching matrix is further optically coupled to a first outgoing fiber and a second outgoing fiber, the second 4X4 switching matrix routing the protection channel of the first incoming OTSn to the second outgoing and routing the working channel of the second incoming OTSn to the first outgoing fiber.

17. The add drop multiplexer of claim 8 wherein the control system configures the matrices to correct for an interruption in both of the incoming OTSns.

18. The add drop multiplexer of claim 17 wherein the control system configures the matrices in accordance with an adjacent node switching scheme.

19. The add drop multiplexer of claim 17 wherein the control system configures the matrices in accordance with a non-adjacent node switching scheme.

20. The add drop multiplexer of claim 8 wherein the control system configures the matrices to correct for a single channel failure.