Patent Grant
7035537
The present patent application is related to fiberoptic networks, and, in particular, to switches for WDM and DWDM network systems.
In WDM (Wavelength Division Multiplexing) fiberoptic networks, optical signals are sent at predetermined wavelengths over optical fibers. Each predetermined wavelength forms a communication channel in the network and the wavelength (or frequency) of the optical signal is used to control the destination of the signal through the network. An advanced version of WDM networks is the DWDM (Dense Wavelength Division Multiplexing) network in which the number of wavelength channel is increased by reducing the channel wavelength separation to 100 GHz, as set by the ITU (International Telecommunications Union). Hence the term, DWDM, is used herein to refer to both WDM and DWDM networks and other fiberoptic networks which rely upon wavelength to define communication channels, unless indicated otherwise.
In networks, including such fiberoptic networks described above, switches or routers are used to select paths for signals through the networks. In fiberoptic networks switches and routers not only direct optical signals from one optical fiber to another but also from one wavelength channel to another. The availability of light paths is critical to the users of a network. One way to provide reliability for a light path within the network is to explicitly provide for a redundant path beforehand. However, this approach does not utilize the bandwidth of the network efficiently, i.e., some of the available network capacity is removed for the backup reserve.
The present invention, on the other hand, is directed toward on-the-fly light path restoration to achieve efficient bandwidth usage and availability at the same time. New paths are quickly rerouted through the network in place of the lost light paths.
The present invention provides for a method of operation in an optical network having a plurality of interconnected nodes with each node capable of selectively switching optical signals in a first wavelength channel in an input fiber to any one of a plurality of wavelength channels and output fibers. The method restores connection between the nodes upon a failure of the network by maintaining at each of the nodes a synchronized database of network connections between the nodes; sending messages to other nodes to initiate restoration operations by a node noticing the failure; and recalculating the network connections around the failure by each node from the synchronized database at each node. Each node performs the recalculation independently.
The present invention also provides for a fiberoptic network having a plurality of interconnected nodes with each node capable of selectively switching optical signals in a first wavelength channel in an input fiber to any one of a plurality of wavelength channels and output fibers. A reserved wavelength channel between the interconnected nodes forms an optical supervisory channel to create a control network useful for network restoration and provisioning operations.
Traditional network restoration techniques utilize a central network controller. When the network controller is notified of a network failure, it may signal provisioning information to one or more nodes in the network to implement alternate routes for circuits affected by the failure. The calculation of the alternate routes may occur before or after the failure.
The present invention seeks to speed up this process by utilizing embedded, distributed control logic in each node. The chief benefit of this distributed approach is a significant saving of signaling times between the nodes of the network.
For example, the line 21 represents a plurality of fiberoptic lines carrying data between the switch nodes 10 and 11. For controlling the operations of the pair of switch nodes (and the other switch nodes), the network reserves a wavelength channel in one or more of the optical fibers to carry control signals. The reserved wavelength channel may be one of the ordinary WDM wavelength channels, or the reserved wavelength channel might be a channel specially created for control signals. For example, a current WDM standard specifies 64 wavelength channels in one optical fiber and the reserved wavelength channel would be the 65th channel placed at the wavelength boundary of the 64 channels. Likewise, the node 10 also is connected to the line 29 which represents a plurality of fiberoptic lines carrying data and control signals between the switch nodes 10 and 14, and so forth.
These reserved wavelength channels form a optical supervisory channel for control and signaling operations for the
Besides the reserved wavelength channels illustrated by the dotted lines 21A, 23A, 25A, 27A, 29A, 30A and 31A between the switch nodes 10–14, an alternative control network can be used, i.e., the Internet. As illustrated in
The particular architecture of the switch nodes 10–14 is illustrated in
The fiberoptic switch of
The tunable channel cards 42 are connected to a switch fabric, which, in conventional network switching terminology, constitutes the portion of a switch which performs the actual switching function. In the case of the
The N×N switch module 43 is formed from N 1-to-N switch elements 45 and N N-to-1 combiners 46. The number 1 output terminals of the switch elements 1-N are connected to the corresponding input terminals 1-N of the first combiner 461. The number 2 output terminals of the switch elements 1-N are connected to the corresponding input terminals 1-N of the second combiner 462. This pattern is repeated for all N output terminals of each switch element 451–45N. Each switch element 45 corresponds to one of the inputs to the N×N switch module 43. To connect any input terminal to a specific output terminal of the switch module 43, the switch element 45 for that input terminal is set to the position for that output terminal. In this manner, signals on any combination of input terminals of the described fiberoptic switch can be sent to any combination of output terminals, with two constraints. Any single input terminal can only be connected to a single output terminal at a time. If multiple input terminals are connected to a single output terminals, the signals on the multiple input terminals must be noninterfering (i.e., the signals must be at different wavelengths/frequencies).
Operationally, to connect an incoming signal on some input fiber 40 at a first wavelength to an output fiber 41 on a different wavelength, two operations must be performed. First, the tunable channel card 42 which is associated with the incoming signal at the first wavelength on the input fiber 40 must be tuned to translate the signal to the correct outgoing wavelength. Also, the switch module 43 associated with that tunable channel card 42 must be configured to send the signal to the correct output fiber 41. These operations of the channel cards 42 and the switch modules 43 are directed by the control unit 47, which contains a processor unit, such as a microprocessor or a microcontroller, and memory 48. The memory 48 includes nonvolatile portions to hold software for restarting switch operations after the system goes down for any reason.
The reconfiguration process can be done fairly quickly. First, the input switch stage (i.e., the switch element 45) is turned off to disconnect the laser source in the tunable channel card 42. In an alternative arrangement, rather than switch elements 45 with off/on functions, off/on switches are placed between the switch elements 45 and the tunable channel cards 42, and the switch for the particular switch element 45 is turned off to disconnected the laser source in the tunable channel card 42. Then the laser is tuned to the new wavelength and the switch elements 45 in the corresponding switch module 43 are set to the correct states for the new configuration and the connection turned back on.
Hence these switches direct optical signals through designated optical fibers 20–31 and through the M wavelength channels in the
The optical supervisory channels in the optical fibers 21, 23, 25, 27, 29–31 carry signaling and control signals between the switch nodes 10–14 for the restoration and provisioning operations. As stated previously, the signaling and control signals are in the form of IP commands through the IP routers 10A–14A. If one or more of the network component fails, network operations must be restored. For example, in the exemplary network of
The optical supervisory channels indicated by the dotted lines 21A, 23A, 25A, 27A, 29A–31A in
To handle these operations, restoration and provisioning software is stored in each switch node 10–14. The software contains both switch node management control software for each switch node and management system software for the whole mesh network. The two software components are divided and the management system software can be placed in the switch node or in a unit separated from the switch node, such as a standalone UNIX/NT box. The software interacts with two management information databases also stored at each switch node. One database is a local management information database which holds information about the switch node and the other is a network management information database which contains the cross-connect provisioning status across the entire mesh network. Only the restoration and provisioning operations can result in a database change. But the network management information database at each switch node must be guaranteed to be consistent across the entire mesh network with the other switch nodes for proper operation of the network. This is carried out by network synchronization.
To guarantee the database on each switch node is synchronized, the network operation is carried out either at every node or none at all. The initiating node of the operations determines whether the operation is successful or not, based on the acknowledgment from each node. The transaction is completed if every node carries out the operation successfully. Otherwise, the initiator sends out an abort transaction message to every node to cancel the operation.
The restoration operation is carried out in the following manner: First, a network restoration initiation message is broadcast through the network by the initiating switch node, i.e., the node noticing the hardware failure. The restoration calculation is done independently by each node at each local management information database upon receiving the restoration message and is coordinated by the initiating switch node. The calculation is based on the network management information database which is synchronized to be consistent at all times with the management information databases at the other nodes.
For the provisioning operations, a provisioning command modifies the cross-connect setup of the network and passes through a multiphase transaction protocol. First, the provisioning command is sent through the entire network to reserve the resource to be provisioned. The switch node issuing the command receives responses from all the switch nodes of the network. If all the responses are affirmative, the commanding node sends a “do-it” command to the entire network to do the actual provisioning operations as commanded. All the switch nodes send responses back to the commanding node as to the success of the provisioning operation and the commanding node either commits or aborts the entire transaction depending upon the responses.
Since a restoration transaction has higher priority than a provisioning transaction, the restoration transaction may or may not pre-empt the provisioning transaction before initiating the restoration. This depends upon whether the database is synchronized or not. At any time there can be no more than two transactions in progress, one provisioning transaction and one restoration transaction. Each transaction has a unique transaction number across the whole network.
To carry out these operations, each switch node has a network operation transaction state machine with the initial state as the Idle state. The switch node also keeps a Next Available Transaction Number (NATN), which is initialized to a default value and is then synchronized once it joins the network. The states are as follows:
The transition from state to another in a switch node is triggered by a message or an event. A triggering message is generated by an operation-initiating switch node for transmission to the other switch nodes. A triggering event results from a user request or acknowledgment from other node.
In the described distributed network approach, the principal objective of the signaling protocol is to disseminate the failure event information to every node in the network as quickly as possible. Hence signaling is used for the failure event information only, and not to cross-connect provisioning information. Broadcast mechanisms are used for signaling, which use pre-provisioned fixed alternate routes through the optical supervisory channels.
Each failure event message is identified by the source node and a node-specific Failure Event Message Number. Each receiving node keeps track of each other node's current Failure Event Message Number. If a duplicate is received, it is ignored and discarded. In the case of a bi-directional fiber cut, two nodes detect the same network failure and each initiate the broadcast signaling. In this case, other nodes in the network must reconcile the two failure event messages as describing the same single event. In order to perform such reconciliation, each node starts a timer upon receipt of a failure event message. If another “similar” failure event message is received before expiration of the timer, then the new message is ignored and discarded.
In order to better utilize the embedded IP (Internet Protocol) routers 10A–14A of each node 10–14, multiple fixed alternate routes are pre-provisioned from each node to every other node in the network. Upon local detection of a network failure, a node transmits a series of IP packets, each containing the failure event message, one for each fixed alternate route to each node. The IP router network (illustrated by the network in
The receiving nodes detect and discard duplicate messages. Hence, after a switch node receives a failure event message, the node participates in the flooding protocol described above. The switch node then releases all the wavelength channel resources (i.e., bandwidth) of the optical circuits available for use by alternate routes, despite the network failure. Then in priority order, the shortest path for each affected circuit is recalculated, using only available, i.e., in service but unused, network resources. The optical circuits are restored in a predetermined prioritized order in this manner.
Each switch node recalculates a new path for each circuit whose active path had traversed the failed link. Each node evaluates the results of the path calculation to determine whether or not that node must execute any new cross-connects. If so, the cross-connects are executed. If not, then that node takes no action, and its participation in the network restoration is completed. All switch nodes perform the identical deterministic calculation, and therefore arrive at the same conclusion.
To determine the alternate routes, the switch nodes use a version of E.W. Djikstra's “Shortest Path First” (SPF) algorithm to route circuits. U.S. patent application Ser. No. 09/896,843, entitled “Improved Shortest Path First Restoration Routing In a Fiberoptic Network,” filed of even date by Peter Abrams and assigned to the present assignee, and which application is incorporated herein by reference, describes the modified implementation of the SPF algorithm in its operation in the exemplary
Therefore, while the description above provides a full and complete disclosure of the preferred embodiments of the present invention, various modifications, alternate constructions, and equivalents will be obvious to those with skill in the art. Thus, the scope of the present invention is limited solely by the metes and bounds of the appended claims.
This patent application claims priority from Provisional Patent Application Nos. 60/215,182 and 60/215,399, both filed Jun. 29, 2000 and are hereby incorporated by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4956835 | Grover | Sep 1990 | A |
| 5396485 | Ohno et al. | Mar 1995 | A |
| 5488501 | Barnsley | Jan 1996 | A |
| 5548639 | Ogura et al. | Aug 1996 | A |
| 5581543 | Natarajan | Dec 1996 | A |
| 5586112 | Tabata | Dec 1996 | A |
| 5592467 | Takeuchi et al. | Jan 1997 | A |
| 5598403 | Tatsuki | Jan 1997 | A |
| 5903370 | Johnson | May 1999 | A |
| 5914798 | Liu | Jun 1999 | A |
| 6073248 | Doshi et al. | Jun 2000 | A |
| 6233072 | Liu et al. | May 2001 | B1 |
| 6671256 | Xiong et al. | Dec 2003 | B1 |
| Number | Date | Country | |
|---|---|---|---|
| 20020041413 A1 | Apr 2002 | US |
| Number | Date | Country | |
|---|---|---|---|
| 60215182 | Jun 2000 | US | |
| 60215399 | Jun 2000 | US |
