CA1291549C - Method and apparatus for self-healing and self-provisioning networks - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
A method and apparatus of restoring communications between a pair of nodes in a network having an arbitrary number of nodes and an arbitrary number of spans interconnecting the nodes, each span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits capable of, but not designated for, transmitting actual com-munications traffic, the method comprising the steps of (a) establishing one or more independent communication paths between the pair of nodes through a series of spare circuits of spans interconnecting the pair of nodes and other interconnected nodes in the network; and (b) redirecting communications traffic intended for one or more failed spans interconnecting the pair of nodes through one or more of the paths.

Description

s~

The present invention relates, in general, to a method and apparatus for rapidly ef~ectlng, in a communicatlons network. the restoration o~ communica-tions be-tween nodes whose interconnecting spans have 5 failed ~or one reason or another. More specifically, the present invention pertains to a method, ~pparatus and distributed control protocol for the real-time reconfiguration of Digital Crossconnect Switches to achieve rerouting of traffic around failures in a 10 telecommunications network or ~or rapid identi~ication and provisioning of new transmission routes between any -two locations in a diverse network.

BACKGROUND OF T~E I~VENTION
Restoration is the re-establishment Oe trunk-bearing carrier groups a~ter loss of all or most o~ the physical transmission facility between two sites, through geographicsl rerouting via redundant network capacity~
Rerouting ~or restoration o~ the physical transport 20 network should not be con~used with the routing of individual calls or with the archltecture of the logicaI
trunking network administered at the DS-1 transport level~ These will remain unchanged by a success~ul restoration at the DS-3 level~
As a cable-based technology, FOTS has proven suscep-tible to ~requent damage due to construction work, lightning strikes, rodent damage, craftsperson error, train derailment, etc~ With the inherent capacity Oe FOTS, and reduced physical route diversity in 30 eiber-based networks, a cable cut can seriously a~fect network blocking because each eiber carries many diverse logical trunk groups~ Structural availability of eiber cable in real networks is reported as low as 96.5% (300 hours/year downtime) at a cost up to $75,0UO US per 35 minute of outage. By comparison, radio transcontinental routes typically do meet requirements of 99.985 % ~or structural availabilit~es without special restoration methods. It therefore seems unavoidable that advanced methods o~ restoration are an essential ad~unct to the ~ide~prQad deployment o~ ~iber networks.
The conventional method ~or restoration o~ a ~iber cable cut is manual rearrangement at passive DSX-3 5 crossconnect panels. The ~equence of patch operations is determined ~rom stored plans or i~ developed at the time o~ restoration. Crews are dispatched flrst to pa~ch the rearrangements and then to proceed with physical fault location and repair. The time to restore 10 traffic averages ~rom 6 to 12 hours.
The current industry plans are that wlth the advent of DCS 3/3 (or DCS 3/1) machines, the above restoration method will be automated through remote control of DCS-3 machines. Some o~ the problems and limitations with the lS planned approach are:
Speed: Althougb centralized control o~ DCS--3 machines will signi~icantly reduce restoration times compared to manual patching, there is little expectation of approaching realtime recon~iguration capabilities 20 with centralized control. Estimates within the industry place restoration times initially around 1 hour, improving to perhaps 10 minutes when the centralized DCS
management systems are mature.
Da~abase Dependency: The centralized approach raises 25 concern about the size, cos-t, complexity and vul-nerability o~ the surveillance and control complex that will be needed ~or transport management. The central-ized approach will be dependent on the ability to maintain a complete, conslstent, and accurate database 30 image of the network over years o~ operation. Eventual-ly a maintflnance change that i3 not immediately and correctly re~lected in the database creates the pos-sibility o~ service-arfecting error during a centrally controlled restoration or recon~iguratlon even-t.
Tra~lc Impact: With centralized control, all calls-in-progress will contlnue to be lost whenever a cable is cut because the outage duration remains much longer than voice call-dropping thresholds and data '?6 5L~

protocol timeouts. This means that with central:lzed control, span protec~ion switchirl~ will continue -to be required in transmiæsion sys-tems to handle single carrier failures with enough speed to avoid call 5 dropping.
Tele~etry Dependency: Centralized con~rol of DCS
machines also requires redundant telemetry arrangements 90 facility cuts will not remove the very communicatlons links over which the central control site is to issue 10 restoration commands. This requires redundant com-munications interfaces on the DCS equipment and special circuit engineering considerations for the operating company.

15 SUMMARY OP TH~ INVENTION
The present invention seeks to provide a method and an apparatus which overcome the disadvantages of current restoration methods and apparatus and more specifically which will rapidly and automatically efi'ect restoration 20 between -two or more nodes in a network whose intercon-necting spans have failed.
In accordance with one aspect o~ the present inven-tion, -there is provlded a method of and apparatus for restoring communications between a pair o~ node~ in a 25 network havlng an arbitrary number of nodes and an arbitrary number o~ spans interconnectlng the nodes, each span havlng working circuits between nodes desig-nated for transmitting actual communications traffic and spare circuits capable of, but not designated for, 30 transmitting actual communications trai'~ic. The method comprises the steps of (a) establishing one or more independent communication paths between the pair of nodes through a s0ries of spare circults of spans interconnecting the pair of nodes and other intercon-35 nected nodes in the network: and (b) redirectingcommunications traf~ic intended i'or one or more failed spans interconnecting the pair of nodes through one or more of the paths.

In accordance wi-th a more speciflc aspect of the present invention, t~ere is provided a restoring apparatus incorporated into each node of a communica-tions network having an arbltrary number of nodes and an S arbitrary number of spans interconnecting the nodes, each span having wor~ing circuits between nodes deslg-nated for transmitting actual communications traffic and spare circuits between nodes capable of, but not designated for, the transmi~sion of actual communica-10 tions traffic. Each node has one or more bi-direc~ional transmission interfaces connected to external transmis-sion lines and to an internal switching matrix. Each transmission interface has circuit means for processing signals received along a receive link and feeding the 15 received signal to the switching matrix and circuit means for processing -transmit signals received from the switching matrix and applying the transmit signal to a transmit link. The restoring apparatus comprises restoration signature detecting circuit means at each 20 the transmission interface for detecting and storing restoration signatures received along the receive llnk, restoration signature transmitting circuit means at each transmission interface for applying transmit restoration signature ~ignals to the transmit link; and control 25 means operatively connected -to the detecting circuit means and the transmitting circuit of each the transmis-sion intereace. The control means i9 responsive to (a) an alarm signal indicative o~ a circuit failure in a span connecting the node in which the control means is 30 disposed ancl an ad~acent node for generating and repeatedly applying predetermined restoration signature signals to the restoration sig}lature transmit-ting circuit maans of one or more of the transmitting interfaces whereby to cause the restoration signals to 35 b0 transmitted along one or more transmitting links to other ad~acent nodes, (b) a restoration ~ignature detected by the detecting circuit for producing modified restoration signatures and applying the modified ~¢~

restoration signatures to the transmitting circuit means o~ each -the transmission :Inter~ace o~ the node whereby to cause the modified restoration signatures -to be transmittecl along the transmitting links -to ad~acent 5 nodes, (c) a modified restoration signature recelved at the detecting circuit means of one oE the tranxmissiorl interfaces ~or generating a complement restoration signat~re signal and applying the complement restoration signature signal to the transmitting circuit means of 10 the one of the transmission interfaces, and (d) a complement restoration signature received at the detecting circuit means of one oi~ the transmission interfaces for either operatively connecting the receive and transmission links of the transmission interface at 15 which the complement restoration signature signal was received with the transmission and receive links, respectively, of the transmixsion inter~ace on which the restoration signature associated to the complement signal was received i~ the control means modified an 20 existing restoration signature, or operatively connect-ing the receive and transmission links of the trans-mission inter~ace at which the complement restoratlon signature slgnal was recelved with the transmission and receive links, respectively, o~ the transmission 25 interface on which communications traffic intended to be transmitted through a ~ailed cirouit between the pair o~
nodes ie the control means generated the original restoration signature as~oclated with the complement restoration signature signal.
In this way, one or more alternate and independent communication paths are established between the pair of nodes through spare circuits oi' spans :Interconnecting the pair o~ nodes and other lnterconnected nodes in the network.
36 Thus, the present invention provides an arrangement whereby the computation of network restoration plans are distributed among the Digital Crossconnect Switches (DCS) processors in a network so that reroutings are computed and executed automatically in realtime without recourse to centralized control or da-tabases.
While not limited thereto, the present invention uses DCS machines in the DS-3 or SONET transport network to 5 create a network-wide reflex reaction which achieves ~acility restoration of complete cable cuts in a very short period of time. Experimentation on network models have shown groups of up to eight DCS machines coor-dinating their actions in accordance with the present 10 invention perform complex restorations of complete cable cuts in less than one second and restoration coverage efficiencies have been found to be equal to the restora-tion plans obtainable by human inspection of the same networks. When a single circuit fails, but -the cable is 15 not cut, the present invention behaves like an Automatic Protection Switching (APS) system, reacting in 100 to 200 msec. The technique is distributed and uses the network entities directly as its database. It is contemplated that the present invention constitutes a 20 simple, high performance realtime assistant to central-lzed administrative control of the transport network thereby removing the demand ~or real-time re~ponse from network operations centers.
In the detailed description which follows, reference 25 is made to the combination of apparatus and method as "Selfhealing" and focuses, for practical application and descriptive purposes, on the application oP Selfhealing to the third level of the North Amerlcan and lnterna-tlonal transmisslon hierarchy. Digital Cross-Connect 30 machines at this level are denoted a~ DCS-3 and thc prlme responslbllity eor transport network restoration resldes at this level. It i9 to be understood that DS-3 may equivalently read "STS-1" because the SelPhealing process ls functlonally ldentical at the STS-l in-tercon-35 nect level of a future SONET network. Only the detallsof restoration signature transport vary.
Although Selfheallng itself operates entirely without central control, it does not replace centrallzed network control. Selfhealing is seen to b0 deployed as a real-time field--asslstant to a central:iz.ed Network Operations Center (NOC). In this way, the overall coordlnatlon, monitoring and administration of the 5 transport network remains under centralized control but -the difficult requirement for near-real-time response from the central control complex i9 removed.
The presen-t invention is primarily of value to manufacturers and telephone network operating companies 10 as a realtime assistant to centralized network opera-tions systems by permitting DCS machines to restore cable cuts without call-dropping, replace span protec-tion switching subsys-tems in transmission equipment, reduce speed and database requirements for centralized 15 network control systems, and reduce the structural availability and redundant capacity requirements of fiber optic networks.
The present invention offers the advantages of replacing span protection switching with network wide 20 automatic protection and unlfying the traditionally separate functions of protection swltching and network restoration. ~urther, it has no database requirements, no dependency on telemetry or availability of the central control, and the potential xpeed to restore 25 entire cable outs without call dropping and the prospect of elimina-ting automatic span protection switching (APS) in transmission equipment.

BRIEF DESCRIPTION OF T~E DRAWI~G5 These and other features ot -the invention will become more apparent from the following descrlption ln which reference is made to the appended drawings wherein:
5 FIGURE 1 shows examples of communication3 networks and the components of such networks -that pertain to the present invention;
~IGUR~ 2 illustrates one of the example networks of PIGURE 1 as having undergone a span cut and shows the general manner in which the present invention achieves distributed self-selected coordination amon~st nodes of the network so as to very rapidly restore the traffic affected by the span cut;
FIGURE 3 schematic illustrates a digital cross-connect machine equipped with the necessary elements so that when each node of the network in FIGURE 1 ls so equipped, the objective o~ a self restoring (or "self healing") network is achieved;
FIGURE ~ schematic illustrates signature reception and transmission circuits added by the preferred embodi-ment of the present lnvention to each transmission line interEace block illustrated in PIGURe 3;
FIGURE 5 illustrates the content and organization of information in th0 slgnature registers shown in PIGURE g according to the preferred embodiment o~ the present invention;
~IGURe 8 illustrates the highest level logical structure Oe the restoration controller according the preferred embodiment Oe the present invention as a Finite State Machine having four states including NORM~I., SENDER, CHOOSER and SETUP-TANDEM states and three ma~or logical processing blocks for the manipu:lation of signatures in a manner such that automatic re~tora-tion is af~'ected;
35 PIGURES 7 to 14 illustrate the manner in which the latter three logic blocks interact with signatures in the preferred embodiment of the present invention;

FIGURE 7 schematically illustrates the principle by which a node adopting the SENDER state initiates the forward signature flooding step o~ the pre~ent invention;
5 ~IGURE 8 schematically illustra-tes the principle manner in wbich a node in the CHOOSER state, responds to signatures arriving at its slte, initiating the reverse-linking step of the present invention i`or one of the several restoration pa~hs that may be re-quired;
FIGURE 9 illustrates the principle manner in which aSENDER node reacts to a signature from the CHOOSER
arriving at its site after completion of forward flooding and operating a crosspoin~ to connect to the restoration path that is known by the method of the present invention to now exist between SENDER and CHOOSER nodes;
FIGURE 10 illustrates the step by which the CHOOSER node determines which crosspoint to operate to complete the re-routing of one of the several signals dis-rupted by a given span cut;
PIGURE 11 schematically illustrates the basic manner in which a node in the TANDEM state rebroadcasts an appropriate new signature arriving at its location;
25 ~IGUR~ lZ ~chematically illustrates the manner in which a node in the TANDEM state reacts to the arrival of a signature initiated by the SENDER node when such signature meets the condit:lons of being a superior precursor signature ~or some existing transmit signature rebroadcast than the existing precursor receive signature assoclated with the rebroadcasted signatures;
~IGUR~ 13 schematically illustrates the manner :In which a node in the TANDEM state reacts to the arrival of a reverse-llnking signature when such signature creates a complement pair of signatures on one port of the node in the Tandem state; and FIGURE 14 schematically illustrates the manner in which a node in the TANDE~ æ-tate reacts to the dlsap-pearance of a received signature at its loaation and also shows, in con~unct:ion with PIG~RE 11, how the TAN~EM node reacts to a signature that does not disappear but changes ~ince such an event i~ trea-ted as the disappearance of the ~irst signature (FIGURE 14) followed by appearance of the second signature (FIGUR~ 11).
DESCRIPTIO~
The first portion oE the following description explains the overall principles and mechani3ms of operation over an entire network and introduces the 15 terminology and related concepts required to appreciate the method. The key objective to be explained is the manner ln which distributed DCS machine~, endowed with signature transmission and reception and certain other hardware including a controller to implement the 20 signature manipulation logic, are enabled to individual-ly and independently act in a manner such that the network as a whole rapidly e~ecutes a coherent restora-tion plan, such plan appearing, prior to the pre~ent invention, to be possible only through centrali~ed 25 global observability and control Oe the network.
The second portion oE the following description relates to the details Oe the specific embodiment oE the invention which has been realized and e~plalns the lower-level details throuKh which the obJects Oe the 30 invention are achieved, these more detailed considera-tions being amenable to description only after the more general de~cription has been completed.

GEN~RAL DESORIPTION
Selehealing i9 a property of a network, not o:E a node, although the oecessary elements for a Selfhealing network reside within the DCS machines in the nodes of the network. A network is endowed, in accordance with 5'~

the present invention, with the property o~ Selfhealing by two elements: a Selehealing logic controller in each DCS-3 machine and transparent signaling circuits, physically associa-ted with each -transmission line 5 inter~ace port (DS-3 or STS-~ ~or example), and a corresponding transmission method for the transparent transportation o~ restoratlon signatures over these same transmission links.

SELPHEALING RESTORA~I~N SIGNAT~RES
The various nodes o~ a network equipped with the apparatus ~or Sel~healing interact through restoration signatures on the links between them. Signatures are signi~icantly dif~erent ~rom messages between processors 15 as might usually be provided ln a multiprocessor or in a packet communications environment. A signature is an attribute Oe a unidirectional transmission link and is physically inseparable ~rom that link. A signature is not addressed to any particular other node and Sel~heal-20 ing nodes do not. directly interchange messages ordir~ctly address each other for co-ordination of in~ormation or control between their controllers.
Rather, the execution o~ each Selfheallng controller af~ects the execution o~ all other nodes only through 25 its influence on the signatures in its vicinity. The Selfhealing cont:roller is in one sense a specialized computer structure ~or processing o~ signatures. Each controller instance a~ects its nelghbors by changes it makes ln the number, content and link association o~ the 30 signatures originated from lts slte in response -to the signatures arriving at its site.
Signatures are physically associated wlth each D5--3 (or STS-1) entlty in the network but are invisible to the tra~fic on those linlcs. Selfhealing controllers 3~ initiate, manipulate, modi~y and terminate signatures in a process that permi-ts each DCS machine in the network to derive local crosspoint operate decisions that, through this invention, automatically coordinate with - ~2 -the similarly derived decisions made at other DCS
mach.ines. The ne-twork-level result i~ restora-tlon, although each node has no knowledge Oe the topology of the network that it is in~ In the present implement-5 ation, restoration signatures are 5-word quantities, described later, repeatedly impressed on-to -the iD-dividual DS-3 signals by one o~ several possible methods.
An analogy that ~ay help in the understanding o~ the lO role of signatures is the ~ollowing: Signatures are to some extent like tokens in a game. In this analogy, Selfhealing works through a formal set of rules for the creation and elimination Oe the tokens of the game. In addition each node seeks a goal that is speci~ied in 15 terms of certain types Oe matched signatures (later called complements), which results in a certain reward:
permission to operate specific crosspoints. ~urthering this analogy, no player (node) in the game knows about the higher level objective that is indirectly achieved 20 (network restoration), although the design of the game is actually directed to achieve this meta-objective which is unknowable to any individual player.

MFTNODS ~OR SIGNATURE TRANSPORT
The requirement eor signature transport is easily met in the SON~T network by reserving one Oe the existing un-allocated signaling eields ~or this purpose.
In DS-3 based networks, s:lgnature transport requ:lres new techniques because there is no built-in means to 30 provide for ~uperv:1sory :1neormation transport in the existing DS-3 format. The only deeined ~ignaling entity, the X-bits, are reserved ror customer use. As a related part Oe this work, the present patentee has researched three techniques for transparent auxiliary 35 channel signaling in the DS-3 rormat. A proposed method that is compatible with both asynchronous DS-3 and SYNT~AN modulates auxiliary information onto the F-bits 5~

of the DS-3 in such a manner that framing performance is only -trivially affec-ted.
A second method that provides a higher auxlliary signaling rate but i9 not SYNTRAN compatible, sub-5 stitutes auxiliary information bits for the dummy bitspresent in subframe 7 of the asynchronous DS-3 format whenever positive pulse stuffing occurs.
A third method exists in circumstances where the C-bits of the cooventional DS-3 signal are liberated for 10 new uses. The third method is based on synchronizatton of the intermediate DS-2 tributaries in ~ump-level M13 multiplex ter~inal equipment to permit ganged stuffing in the M23 stage.
Any or all of these signature transport strategies 15 could be used for Selfhealing in DS-3 networks. Each of the methods requires a DS-3 F-bi-t framing function but relatively little additional circuitry. DS-3 ~l-bit framing i8 now a fairly standard VLSI circuit function likely to be provided in the port cards of most or all 20 DCS-3. DCS 3/1 de~igns must perform DS-3 framing for DS-1 access. High performance DCS 3/3 designs do DS-3 framing for performance monitoring reasons. Therefore, none of these methods of signature transport ln DS-3 networks is expected to create a ma~or obstacle to 25 hardware support of Selfhealing in new DCS-3 products.

THE SELFHEALINQ LOGIC CONTROLLER
The Selfhealing controiler is implemented as a state machine, shown at its highest level in ~IG~R8 fi, and :In 30 accordance wlth the Pascal language Protocol Specifica--tion which appears in Appendix A. The Selfhealing controller has flve states and three ma~or control logic functions, the general behaviour of whlch will become apparent from the following first level description of 35 the manner in which the Selfhealing control logic interacts with signatures in the oetwork to cause the overall result of automatic distributed co-operative restoration.

?J~

A Selfhealing networ~ restoration event can be analy~ed as having two ma~or conceptual phases: 1) a network signature flood.ing wave, and 21 a reverse signature linking sequence.

~ etwork Signature Plooding: A Sel~healing action begins when a cable cut (or equivalent emergency~ occurs on a transmission span within the network. Whenever this happens, existing transmis~ion monitoring equipment 10 raises central o~fice maiDtenance alarms within ap-proxlmately 10 milliseconds of the loss o~ transmission integrity. In this invention, the DCS machirles in the two affected end nodes are arranged to be immediately informed o~ any such tra~ic-affecting ma~or alarms.
15 The alarms may be generated by the transmission terminal equipment and wired to the DCS or may be generated by appropriate circuitry in the port cards of the DCS
machines depending on the CØ equipmerlt arrangements.
In either case, reliable detection of the fault with 20 adequate persistence checking is an established aspect of transmission system design, not part o~ the present invention, except that these alarms must be communicated to the DCS machine at each node in the fas-test possible manner, preferably by direct electronic connection to 25 the primary alarm detection circuits. It is to be noted that, given a network with Sel~healing proper-ties, the artificial creation of pseudo-faults i9 of interest as a means to exploit the basic routing mechan:lsm as an advanced real-time provisioning tool. A disc~lssion o~
30 this varlation will ~ollow.
The occurrence o~ a tra~fic-a~ecting transmission alarm causes the a~fected DCS to give pr:lority to -the Sel~healing controller as opposed to the normal opera-tional controller shown in FIGUR~ 3. In the most 35 ~rsquent and practical case o~ one network span-~ailure at a time, the Selfhealing controller is initially in the Normal state. Multiple simultaneous faults can be handled but are dis¢ussed later. In the Normal state, the Selfhealin~ controller first identi~ies the number and port IDIs) of -the ~ailed transmission llnks. The controller then reads -the last valid contents o~ a receive signature register on the afrected port(s), 5 learning the identlty (ID) of the node to which connec-tlvity has been lost, the remote-failure node. This is the pre~erred form of Selfhealing ln that each node maintains strictly zero knowledge except its own name.
All information that is needed arrives at each node on 10 the links connected to it. Rearrangements in the network are thereby self-updating so no node can have outdated information. A more conservative approach is to store and maintain neighbor node ID tables in each DCS. Selfhealing can also work this way, but with 15 additional administrative requirements.
When the Selfhealing controller recognizes the alarm and has read the receive signature registers on the failed port(s), it then does a simple ordinal-rank test on the remote-~ailure node ID with respect to its own ID
20 to determine whether to act as 'SENDER' or 'C~OOSER' in the subsequent restoration event. The outcome of this test is arbitrary but it guarantees that one node will adopt the SENDER role and the other becomes C~IOOSER.
Every node controller is capable of either role and will 25 adopt one or the other depending on the arbitrary name-rank o~ the two nodes involved by any given span failure. Any arbitration rule conducted independently at each node with the limited information available to nodes and which ensures the above complementary role 30 adoption will sufflce for th:ls aspect of the invent.ion.
The node adopt:lng the SENDER role immediately triggers what is called the Porward S:lgnature Plooding wave by broadcastlng appropriately indexed restoration signatures on some or all spare DS-3 transmission links 35 leaving the location of that DCS, wherever those links go in the network. In milliseconds, these signa-tures then appear in the signature-receiving circuits of the DS-3 inter~aces o~ neighboring DCS machines and cause 5~

those DCS machines to activate their respective Self--healing cnntroller.
~ hen the Selfhealing con-troller i~ activa-ted by receipt of a restoration signature on a normal spare 5 circui~, rather than by an alarm on a working circui-t, the controller enters the Setup-Tandem state, also called simply the TANDEM state. Selfhealing controllers in the Setup-Tandem state execute the Selfhealing Tandem Logic block of FISURE 6, the main effect of which is the 10 selective rebroadcast Oe incoming signatures on all spare links ~rom that node location. This activa~es yet more DCS nodes into the Tandem state and 90 on so that DCS nodes throughout a certain range o~ the SENDER are rapidly alerted into the Tandem state and a large number 15 (possibly all) o~ the spare links of the network within the affected range have been impressed or modulated with the appropriate set o~ unique signatures. The extent o~
the influence of the forward flooding signature wave i8 controlled by a maximum repeat (MR) parameter which sets 20 the maximum range (in spans) from the SENDER node ~ithin which Selfhealing is allowed to range. Signatures exceeding the MR are not propagate~ eurther. An advanced ~trategy is -to let the MR lncrease as a function of time and of the current degree o~ success 25 after the initial fault. ~nother var:lation is to make the MR a table-function of the node-to-node relation requiring restoration. Use o~ the range-o~-influence limiting MR mechanlsm is a practical, not an essential, consideration. This invention also functions safely and 30 reliably without any range limitation mechanism.
Therefore, ln the manner ~ust described, the ~:lrst reaction o~ the Selfhealing network to a facility ~ailure i8 -the determination of SENDER and CHOOSER nodes for the pa~tlcular failed span, a wave of restoration 35 signatures radiating from the SENDER, identli'ying and allocating a signature to existing spare links (in a detailed manner which follows), and a collection o~ DCS
machines Oe the network which are alerted to the Tandem state and ready to eur-ther help in thq co-operative re-routing ef~`or-t.
As can be appreciated, an important eventual e~:~ect of the ~orward flooding process is that i~ there are any 6 possible routings that could be con~-truc-ted by the series concatenation of spare links between SEND~R and CHOOS~R in the given network within the allowed range of influence, then one or more si~natures o~ the forward flooding wave signatures will arrive at the CHOOSER
10 node. If this occurs, the CHOOSER node triggers the reverse signature linking sequence. If no signatures ever reach the CEIOOSER node, then it can be shown that no method, whether using centralized control and overall network knowledge or not, can restore the given ~ault in 15 the given network.

Reverse Signature Linking: In the cases of realistic interest in telecommunications networks, which are by necessity designed with adequate route and capacity 20 redundancy, one or more ~orward flooding signatures eventually, -through the above process, reach the C~IOOSE~
node. This implies there is at least one potential series o~ spare links between DCS machine~ related backwards in a cause-e~ect chain anchored at the 25 SENDER, but it does not yet mean that a restoration path has been ~ound. A signature that arrlves at the CHOOSER
doe~ not uniquely ident:lfy a route or even imply the existence o~ a unique route for every signature, nor does a number o~ signatures arrlving equal to the number 30 of lost circuits imply a number o~ routes can be ~ound that equal the restoratlon requirement. The problem at this stage i8 to select and actualize or construct, and then uniquely :Identi~y end-to-end, the required number of new routes between SENDER and CHOOSER from the 36 signatures now existing on spare links throughout the network. It is part of tbis invention that this ~unction is accomplished in a distributed co-operative manner as ~ollows.

WheIl certain ~orward ~loodlng signatures arrive at -the CHOOSER node, a reverse-linking prc)cess ls trlggered whicil will trace out one o~ the potential paths as desirefl, throu~h a number o~ co-operatlng Tandem nodes.
5 To lnitiate the reverse-linking mechanism so as to create one o:~ the reqIllred rerouting paths, the C~IOOSER
applies a complementary signature (to be described in detail) on a link having a preferred arriving signature.
The CHOOSER node emits no signatures other than in 10 response to certain forward flooding signatures that arrlve at its Qite.
When this rever~e-linking slgnature arrlves at a node that is in the Tandem state, the Tandem-state Selfheal-ing controller acts accordlng to the Tandem Logic bloc~
lS in ~IGURE ~. The detailed behaviour of the TANDEM node in these circumstances ls described later. Suffice lt to say now that the net e~fect of this is the selective deletion Or certain forward ~looding signatures ~rom the Tandem node site, the posslble operation of a speci~ic 20 DCS matrlx crosspoint and the manipulation and retransmission o~ the complementary slgnature on the transmit dlrection of the DS-3 on which the appropriate ~orward ~looding signature is present.
When, through one or more tandem nodes acting as 26 above, a complementary reverse linking signature arrives back at the SENDER, a comp.lete bi-direct:lonal restora-tion circuit i9 known to have been established between SENDER and CHOOSER. ~lthough this new c:lrcuit can be routed through a number o~ cooperatlng DCS (up to the MR
30 limit) which have already operated the crosspoints requlred along this new route. The SeNDER node does not know the routing o~ this new path, but can deduce its span length (~rom slgnature-borne in~ormation) and is assured by the method revealed here that the other end 35 o~ the new path is indeed connected to the CHOOSER node, i.e. the node to which connec~ivlty has been lost.
The SENDER then operates local cro~spolnts to substitute this path ~or a certain one o~ the ~ailed ~?.,~

circuits and suspends certain of its original signature broadcasts. Thc Sel~healing control loglc ensur~s that SEND~R and CHOOSER both substitute this route for the same one of the many ~S-3 enti-ties that may be managed 5 in a single restoration. The overall mechanism des-cribed above for achieving a single rerouting path actually occurs in parallel for a number o~ reroutings in a Selfhealing network until no further restoration is possible or all traffic on lost circuits has been 10 restored. It is part o~ the art of this invention that what occurs as described above to find one rerouting also succeeds for a number of reroutings being sought simuitaneously and in parallel using the ~ame cooperat-ing Tandem nodes. The overall complexity of Selfhealing 15 network behaviour when simultaneously restoring a number of lost circuits in a realistic network topology is very high. Consequently the way this invention has neces-sarily proceeded was to focus on achieving the exact specification of rules for signature manipulation by 20 Selfhealing DCS nodes using empirical methods to determine when the desired overall network behaviour is achieved. In the remainder of -this description, the basic rules of behavlour implemented by each node such that the desired network-level properties are exhibited 25 are described without any further attempt to deal with the overall complexlty of events at even a sin~le Tandem node involved in a Sel~healing event. With the detailed signature processing loglc o~ the preferred embodiment described later, the following network properties are 30 achieved.

PROPERTIES O~ TH~ SEL~HEALING ~ECEiANISM
Parallells~: The forward-elooding ancl reverse-link-ing sequences race in parallel over dif.ferent routes, 35 limited only by the speed of detection and reaction of the Selfhealing controllers to signatures. Selection of minimum distance paths and correct ordering of rees-tablished ci~cuits are m~na~ed by the SEND~R and CHOOSER
states executing in the end nodes of the ae~eoted span.
Equilibriu~ Conditlon~: If 100% coverage is possible for the topology and span provisioning presen~ then all 5 signatures initiated by the SENDER are eventually matched by complementary signat~res from the CIIOOS~R, or they were rescinded by the SENDER as the needed routes were collected. ~s equilibrium approaches, DCS machines that are not required in the final restoration pattern 10 soon have no incoming signatures and return to Normal.
Those DCS machines with matched forward and complement signatures at their location enter the STABLE-TANDE~
STATE and relinquish control to their Operating System, having activated one or more crosspoints. The~e 15 crosspoints remain closed until the Selfhealing con-troller is later re-activated by another state change in the signature receiving circuit and finds that -the signature anchoring a given crosspoiDt has now been removed.
I~ only partial restoration is possible within the given network and MR range, some unsatisfied signature broadcasts persist in the network until a SENDER-timeout clears unmatched signature initiations from the SENDER
site, acknowledging that no ~urther restoration is 26 possible. This causes release of all other unmatched signa-tures in the network and the resultant flnal state is stable as above except that eewer than 100% of circuits are restor0d.
Single Cirouit Fallure~: Automat:lc Protection 30 Switching (APS) subsystems are traditionally provided in transmi~s:lon systems to pro-tect tra.efic when a single circuit ~alls whlle the cable remains intact. This ~unction is automatically perrormed by Selfhealing without special considerations. Behaviour is the same 35 as tlIat described above with the addition that in the APS situation there is at least one spare circuit on the same ~pan as the circuit that has fa:iled. Sel~healing always makes pre~eren-tial use o~ such circuits, if present. Restoration is very fast in this case because no tandem DCS is involved.
Reporting to Central Control: Immediately after a Selfhealing reaction neither S~NDER, CHOOSER nor the 5 central site has knowledge of the new routings syn-thesized by the network. However, both SENDER and CHOOSER know that throu~h the cooperation of unseen crossconnects, the restored DS-3 entities do have -their far ends connected to the desired node, and the emergen-10 cy is over. Each DCS that was involved reports itsparticipation in a Selfhealing event to its central control si-te over its telemetry link. Indeed these very telemetry links may now be routed through the restora-tion path produced by Sel~healing.
In non-critical ~ime, the NOC can then construct and verify a network image of the failed span and the restoration pattern that was deployed by Selfhealing.
The NOC can now override or alter the initial realtlme restoration pattern for any o~ lts own reasons and 20 dlspatch -the repalr crews. The NOC can also control the subsequent reverslon to -the repalred circuits by commanding the SEND~R node to cancel all restoration signatures or cancel them selectively. Removal o~ these anchorlng slgnatures by the SENDER briefly re-invokes 25 the Selfhealing controllers at co-opera-ting nodes and the C}IOOS~R node. These Selfhealing controllers then release the selected crosspoints and return to the Normal state.
An lnterestlllg re-use of Selfheallng technology is 30 available to the NOC in the time aeter the restoration.
If the NOC temporarily instructs all DCS machines to disable their Sel~healing controllers, the NOC can then directly control the transmlssion oE audit slgnatures between DCS machines to verify the logical connectivity 35 (and error performance) o~ the repairs done ~y field crews before returning the repaired circuits to service.
Tra~ic Sa~ety: The above mechanism required no stored knowledge of network topology at any DCS and operated only on spare circuits. Each Selfhealing controller knows only the name o-f :Lts host node.
Selfheallng is therefore insensitive to rearrangements, growth, out-of-service conditions, e-tc.and is safe from S the conventional databa~e consistency problem without resorting to data-locking mechanisms. The real ne-twork elements, in their exact configuration at the time of a failure, is the database used by Selfhealing.

10 DESCRIPTION OF PREFERR~D E~BO~IM~NT
With reference to FIGUR~ 1, the present invention is a method and apparatus for improving the availability (the fraction of all time during which communication can be achieved between any two nodes of the network) of 15 communications in networks, generally designated by numeral 10, such as shown in ~IGURE 1. Such networks are comprised of an arbitrarily large number of nodes 12 and an arbitrarily large number of spans 14 ~ntercon-necting those nodes. Each node is connected by a 20 minimum of -two spans (physically separate routes, each having one or more circuits) to other nodes in the network. Each span 14 of such a network i9 comprised of an even number of unidirectional transmission links 16 and each link is comprised of a traffic-carrying 25 communications payload signal plus a field for the conveyance of a signature as added by this invention.
All links are associated into f:l~ed pairs inclllding a receive link 1~ and a transmit link 20, one in each direction between the end nodes. Each pair is said to 30 form a circuit between nodes.
The nodes of FIGURE 1 are the central offices of the inter-city or inter-off:lce metropolltan telecommunica-tions network. In this context, each "circuit" o~ the transport network in FIGURE 1 is actually a carrier 35 group which corresponds to a large number of individual telephone connections (e.g. 672 at DS-3).
The circuits comprising each span are in one of -two possible states called Working or Spare. A WORKING

P~ 5~

CIRCUIT is one that has been designated for the trans-mission of ac~ual communlcations -traffic, such as individual telephoDe calls or 64 Kb/s data connections established by the voice level switching machines in the 5 transmission hierarchy. Conventional telephone switches that handle individual calls are not shown in FIGURE 1, only the Digital Crossconnect Switching (DCS) machines of the transport network are shown. The difference between these two is that the volce switches or "traffic 10 switches" set up individual calls through a transport network that is relatively static but whose transmission integrity and s.izing to avoid call-blocking at the traffic switches, is managed by the DCS or "transport"
switches. A SPARE CIRCUlT is one that is in all 15 transmission terms identical to a working circuit, but is not available for traffic carrying call set-ups by the voice switches, ei-ther because it is deliberately set aside for redundancy in the case of failure of a working circuit, or because it is simply present due to 20 the natural provisioning module sizes of -transmission equipment. Both of these reasons are common in current telecommunications networks; the latter e~fect is particularly true in fiber optic networks where the provisioning ocours in very large capacity increments.
26 Economics often dicta-te installation of a 135 ~b/~ (3 DS-3s on one flber) or even a 565 Mb/s (12 DS-3s on one fiber) system even if only a fraction of the total capacity i9 required at the t:lme of installation.
The present invention comprises a method and apparatus 30 which is placed at the nodes o~ a commun:lcations network of the type in ~IGUR~ 1 ln which DIGITAL CROSSCONNECT
SWITC~IES (DCS) machines are installed at each node of the network. The invention is embodied at the DCS
machines where it produces the desired effect tha-t 35 trafPic between the nodes A, B, C, D, E, E etc. of the network can be nearly instantaneously restored in the event of a failure of a link or all o-f the links on a span, by rerouting through distributed switching - ~4 -operations at other nodes of the network, wikhout the need o~ any centrallzed or global coordination.
FIGURE ~ schematicall~ illu~trates the form o~ result that the presen-t inven-tion achieves. In the network o~
6 FIGURE 2, each node A through X has a D~S machine and embodieg the presen~ invention and there happens -to be one spare circuit per span. The example shows a span cut severing four working circuits between nodes N and U
and shows how nodes X,W,V,T,P,O,J,K were self-selected 10 and coordinated in the completely distrlbuted manner of this invention to provide complete restoration of the four lost circuits between N and IJ, through the follow-ing re-routings outlined in bold: N-W-T-U; N-X-V-U;
N-0-U; N-K-J-P-U. In this ac-tual case, restoration was lS achieved in a total time o~ 300 milliseconds, as compared to minu-tes to hours curren-tly anticipated by the industry for centralized telemetry and control o~
new DCS machines. This is a simple example. Much more complicated re-routings are o~ten obtained involving 20 nodes cooperating in many (not ~ust one as above) o~ the individual re-routings. Such complicated routings are less obvious to the human eye than the choserl example, but in all ca~es observed, are the most e~icient that could be achieved.
In this descrlptlon, nodes lnvolved as helpers in the co-operative restoration e~ort, such as X or W in FIGURE 2, are called TANDEM nodes. Nodes a~ected directly by the span ~ailure are in general called the ~ault nodes but, as will be seen later, they are treated 30 separately and specifically referred to as SENDER or C~OOSeR nodes in the descrlption o~ the invention.
~ IGURE 3 expand~ the vlew o~ the DCS machines with modi~icat.tons required by thls lnventlon, placed at every node o~ the networks in ~IGURES 1 and 2. Wi-th 35 reference to FlGURe 3, it can be seen that the creatlon o~ a Sel~healing network requires two additional elements in the crossconnect machines of the network.
In ~IGVRE 3, components 30, 32, 34, 3~, 38, 40, 42 and !5~

~ comprise the normal general embodtment of a crosscon-nect machine, or the purpoæes that are essential here.
The dashed outer boundary in FIGURE 3 represents nodes.
The known crossconnect machine s-tructure ls comprised of 5 bi-directional transmission interfaces 30, which are connected to external -transmission lines 16, which may be coaxial cable, fiber, or radio, for example, and connected internally to the switching matrix 32 of the DCS via paths 34, 36. The switching matrix is operated 10 by a crosspoint operator means 38 which usually com-prises hardware and software. As known in the art, the crossconnect switching matrix and all interface cards are directly monitored and controlled exclusively by the DCS operating system 40 and matrix reconfigurations are 15 only perform0d through external command and telemetry links by a remote administrative centre often called ~OC
(Network Operations Center). The current art has no provision for any form of signaling on traf~ic-carrying tranemission slgnals 16 to be detected and used to 20 control the switching matrix directly.
One aspect of the present invention is that it specifically adds means 50 to detect a specialized form of signaling, called signatures, added transparently to the carrier signals of the network and provides a method 25 for directly reconfiguring the switching matrix, without remote control, in accordance with a certain logical method implemented by the emergency restoration or '5elfhealing' contro:ller 70.
This invention adds to the known art o~ crossconnect 30 machines, signature recelving and transmitting circuit means 50 on each transmission interface card 30 and a special independent controller 70 for processing receive signaturss, controlling transmit signatures and recog-nizing signature determined conditions under which the 35 Selfhealing controller will request operation of a specific crosspoint in switching matrix 32 through shared access to the Crosspoint Controller 38.

FIGVRE 4 is an expanded view oP the slgnature transmit and receive ci~cuitry shown as 50 in ~IGUR~ 3.
FIGUR~ 4 is pre~ented for consideration in conJunction with ~IGURE 5 which details the logical content of -the 5 signa-ture receiving and transmitting registers and associated status ports shown only as registers in FI~URE 4.

RECEIVE SI~NATURE CIRCUIT FUNCTION
With reference to ~IGURE 4, the existing receive propagation path 18 and -the existing receive signal processlng interEace card 30 are shown as they were in FIGURE 3 and item 50 of FIGURE 3 is exploded for closer view. The existing receive signal processing circuitry 15 ~2 is connected to the new circuits only in that a copy o~ the composite signal or a copy containing only the logical binary content of the received traf~ic signal plus overhead is provided as required -to the signature detector circuit 52. Signature detector circuit 52 20 processes the received signal in a manner such that the particular bits or other attribute(s) of the signal used for transparent conveyance of the signature information is recovere~ so as to identi~y the individual signaling elements of the signature on that signal, if any. Two 25 illustrative examples of this process are the ~ollowing;
(a) detection o~ F-bit pattern errors as in Grover Canadian Patent application Serial No. 53~, 090 entitled "Method an~ Apparatus ~or Frame-Bit Modulation and Demodulation o~ the DS-3 Signal" and (b) in the SONET
30 signal ~ormat, stripping Oe~ the designated signaling overhead bytes. It is ~:lthin the scope o~ this inven-tion that any number o~ schemes eor modulation and detection of transparent signature transport methods are possible ~or implementation of the invention in trans-35 port networks using di~ferent transport signal ~ormatssuch as DS-1, DS-3, Syntran, Sonet, or CEPT Eormats.
Having detected the individual signaling elements of the signature, signature detector circuit 52 feeds its . $` b,~

output into a signature reception register 6~. Each -time a complete and valifl signature ls recognized by a validity and logic control circuit 56 to be present in th~ signa-ture reception register 5~ and tha-t signature 5 is also seen by the logic control circuit 56 to be different in from the current contents of a signature storag~ register 5~, then the controller 56 will cause the transfer o~ the newly received signature in signa-ture reception register 54 to receive signature output lO register 58, where it will s-tay until ano-ther valld but dlfferent signature appears on the link 18.
Once the new signature is in the receive signature output register, bus interface 60 connected to Sel~heal-ing logic controller 70 (~IGURE 3) raises an internal 15 flag indicating that a new signature has been received and, when appropriate, communicates the signature to -the controller over the bus and clears the 'new signature' flag once -the controller has read the signature.
Thereafter, the signature remains addressable ~or 20 reference by the controller as may be needed in execut-ing its control logic sequences.
Many means for accomplishing the equivalent peri-pheral-to-controller transeer function are known includlng direct connections between each DCS port 25 signature circuit and the controller or there may be a shared interrupt line followed by polling or a vectored interrupt arrangement, etc. In all Oe these schemes, however, it is an aspect o~ th:1s invention that the controller needs no memory of its own since all signa-30 tures and status data which it needs tu implement :1tsfunction are retained as a distributed memory on each in-t0rface port card of the UCS machine. It is an essentlal aspect, however, that either the controller polls continually for new signatures or each IIQW
35 signature i8 explicitly brought to the controllers attention by an interrupt because the Selfheali.ng control logic is implemented as an event-driven finite state machine and even-ts are determined in terms o~

signature appearances, changes and disappearar1ce~ (to be described further).
In the pr~ferred embodiment, when bus in-terf~ce 60 i8 alerted that a new signature has been received, the bus 5 interface logic completely handles the subsequent transfer of the new signature contents and the iden tification of the respective DCS interface port along with the status bits to the Selfhealing controller 70 in FIGURE 3 by known bus interfacing and transaction lO techniques.
With reference to ~IGURE 5, the contents of any valid signature received off a transmission interface are shown and will be discussed in detail. Suffice it -to say now that checkbit fields for any number o~ known 15 error checking and/or error detecting schemes can be appended to the signature data field shown in FIGURE 5, and such circuitry would also be provided in signature reception register 54 or control logic 56 of FIGUR~ 4 without changing the nature of this invention.
TRANS~IT SIGNATURE CIRCUIT PUNCTION
The signature receive and transmit circuit 50 includes a transmit signature stnrage register 80 whlch can be loaded by Selfhealing controller qO over bus 82 25 and bus interface 60. This provides a means by which the Selfhealing controller can apply a desired transmit signature to any de~lred transmiss:10n path 20 as may be appropriate to implement the logic of the Selfhealing method of the present .lnvent:lon embo~ied in controller 30 70 in ~IGURE 3. Through control logic 84, the bus lnter~ace can load a new signature into Reg:ls-ter 80 and then, once loaded, cause the repeated circulation of the contents of the reg:1ster 90 that, by means of a transmit signature modulation circuit 86, the transmit signature 35 in register 80 is indefinitely repeated on the outgoing transmission link 20 Also shown in ~IGUR~ 4 is a Port Statu~ Register 90 which store~ certain information used by controller 70 in the preferred embodiment. Th0 contents of the port status register can be read by controller 70 and, in the case of an alarm, the alarm-detection hardware 1s arranged to activate bus interface 60 ~o forward -the 5 contents portnumber register 92 and port status regis-ter 90 to controller 70 without delay. Port status register 90 contains elements ~00, 102, 104 and 106 shown in FIGURE 5. ALARM lOO ~FI~URe 5) is a single bit that is set if the corresponding DCS transmission interface 10 experiences a receive loss of signal, a bit error rate degradation, or a loss of timing, etc. or, if external transmission terminal equipment in the same building has detected an alarm on the corresponding channel and this information is connected over to the respective DCS
l5 in-terface port. Those skilled in the art of transmis-sion equipment and inter~ace design will appreciate that there are many techniques and criteria for the rapid determination of loss oi' transmission integrity. The alarm bit is set by hardware and may be reset by the DCS
20 controller after acknowledgement.
Also present in the Port Sta-tus Register i9 a SPARE
bit 102 (FIGURE 5). This status bit is written and maintained by DCS operating System ~0 under conditions o~ normal operation. It indicates whether the given 25 port interface is at any given time in traffic-bearing use or i8 in an equipped-but-idle configuration. If SPARE is true and ALARM is false, then the given transmission lnterface port is available for u9e in the Selrhealing method. If SPARE i9 false (the interrace is 30 carrying trar~ic) and ALARM becomes true, then Selfheal-ing will work to restore the lost traffic by network rerouting on this and all similar working links afeected by the ~ailure event.
The CO~NECTED bit lO~ is not essent:lal but is a 35 single blt which logs whether this :Interface port, whether SPARE or not, is actually connected through ~he matrix to any other interface port forming a path through the DCS machine. If CONNECTED i9 true, then the 5~

eield denoted ~SSOC-PORT-ID 106 contains the number of the port to which this por-t is presently conrlected through the matrix.
If CONNECTED is false, then ASSOC-PORT--ID may elther 5 contain a nul indicator or it may store information used by the Selfhealing controller to assist in the faster manipulation of signatures according to the Sel~healing signa-ture processing method to be described. Specifi-cally, when a DCS rebroadcasts a signature arriving on a lO certain port to a number of other ports as in the forward flooding wave, the ASSOC-PORT- register Oe each port that is transmi-tting a rebroadcast signature stores the port number where the precursor signature for these repeated signatures is found. Depending on circums-tan-15 ces, ASSOC-PORT registers and CONNECTED status bits can be written either by the normal DCS or Selfhealing DCS
controllers.

CONT~NTS 0~ A SEL~ALING SIGNATURE
The following de~cription refers to ~IGURE ~. There is only one basic format of restoration signature in this invention, but it is interpreted slightly dif-~erently depending on whether lt is a "transmit signa-ture" or a "receive signature". Every transmit signa-25 ture becomes a receive signature at the opposite end ofthe link on which it is transmitted. Eaoh DCS inter~ace port 30 has provision for on0 transmit signature and one receive signature. The field NID ~Node IDentifier) llO
is written to a -transmit signature by the node sending 30 that signature and contains the network-wide identifier o~ the nodq originating the slgnature. This NID field appears in the NID field 120 of the corresponding receive signature at an ad~acent node. The NID fl01d ls used ln Selfhqallng so that each DCS machlne can 35 recognize the grouping of links arrivlng at its site into logical spans to the same ad~acent node by as-sociating all links wlth equal receive signature NID

~L?.,~

fields intD a loglcal span for Selfhealing control logic purposes.
~ n al-ternative i~plementation is to store data at each DCS about the facility spans terminating at its 5 site in which case NID's are no-t needed. }lowever, -~he NID-based embodiment is preferred because it is consis-tent w.ith an obJective of the presen-t invention that all in~ormation needed by each node to per~orm Selfhealing comes to that node through the network connections 10 themsel~es, thereby continually being up-to-date and eliminating the significant practical problems of maintaining accurate real-time distributed network configuration databases at each node.
The other fields in a Selfhealing signature appear 15 both in the transmit an~ receive signature registers and, altho~lgh numerically different between transmit and receiv~e sides in any given case, these fields are functionally identical. The SOURCE field of any Tx or Xx signature, 112, 122, respectively, and the cor-20 responding TARGET fields 114, 124 identify either theSENDER or CHOOSER node in a Selfhealing event. SENDER
and CHOOSER nodes are the only two nodes which create brand new signatures in the network (TANDEM nodes only rebroadcast and re-index existing primary signatures but 25 never change Source, Target or Index fields). When the SENDER initially floods the network as described earlier, it applies its own network-wide node ldentifier to the SOURCE field (as well as NID) Oe each eorward flooding signature that it initiates and places -the 30 network identifier of the CHOOSER node (the node to which connectivity has been lost, viewed by the SENDER) into the TARGET ~ield. The node identifier of the C~1009ER node ls known by the SENDER because this i~
latched in the NID field of the receive signature 35 register on the working ports that were ai'fected by the span failure (when an alarm occurs, the last valid receive signature is held by the circuit of PIGU~E 4).

At all other network nodes actlng as tand~ms in the Selfhealing process, SOURCE and TARGEI' lnformation fields are never altered, but are used by each site to distingu:ish betweerl possible independent slmultar~eous 5 failures, i.e., a signature arriving at any third site with a g:iven SOURCE, TARG~T pair is uni~uely identified as pertaining to one particular fault in the network.
Although single isolated faults are the overwhelming case, this feature of the restoration signatures permits 10 Selfhealing to act simultaneously on a number of faul-ts, without hazardous confusion. ~t should be noted that if only one fault at a time is somehow guaranteed or is the deliberate limitation oP the design objective, then SOURCE and TARGET fields can be eliminated and replaced 15 by a single bit that indicates only whether a given signature is "forward" (initiated by SENDER) or "reverse" (initiated by CHOOSER).
When SOIJRCE and TARG~T ~ields are used, SOURCE
becomes the CHOOSER node ID on those signatures init-20 iated (crflated) by the CHOOSER in response to -the forward flooding wave of signatures arriving at its site. As above, the CHOOSER similarly knows the node ID
of the node to which it has suffered the connectivity loss, and it writes th:ls node ID into the TARGET field 25 of reverse-linking signatures that are initiated from its ~ite. In summary, therefore, SOURCE and TARGET
signature fields serve to separately identify individual simul-taneous network faults if they occur and, through the following relations, forward ~looding and reverse 30 linking s:lgnatures are distingulshed (because this is essential for all three Sel~healing states to act correctly) as follows: A forward flooding signature has SOURCE = (SENDER node), TARGET = (C~IOOSER node) and a Reverse linking signature, i.e. from the CHOOSER, has 36 SOllRCE = (C~IOOSER node) and TARGET = (SENDER node).
The INDEX field 116, 126 of a restoration signature is an arbitrary serial number given to each original forward flooding signature initiated by the SENDER node at the commencement of the Selfhealing distributed res-toration procesx. INDEX is never altered by any other node. It wlll be s&en that, ln con~unction with the pattern o~ original signa-ture flooding done by the 5 SEND~R and the subsequent reaction to signatures at Tandem nodes, the INDEX field has the effect of managin~
the complex parallelism of the Selfhealing route-~inding method so that independent end-to-end parallel routes are found rather than routes comprised of concentrated lO parallel and serial segments. Another way to think of the role of INDEX is that the distributed Selfhealing mechanism acts independently, although in parallel, to try to create one complete end-to-end path or no path for each different IND~X value issued by the SENDER. If 15 the number o~ di~ferent IWDEX values issued by the SENDER in the original floodlng sequence is greater than the number oP transmission paths lost and the network is proceeding to synthesize an excess number o~ paths, the SENDER node simply suspends the excess signature 20 initiations to obtain only those number of paths needcd from the network. When a C~OOSER creates a new rever-se~linking signature in response to a given forward flooding signature, it uses the INDEX value oP the corresponding forward signature.
26 The remaining signature field, REPEAT 118, 128 provides a mechanism for controlling the range of signature propagation in a Selfhealing action. In the preferred embodiment of the present invention, it is simply an integer value assigned to 1 in any signature 30 when created by SENDER or CHOOSRR. Every subsequent Tandem node that sees any signature increments the REPEAT field oi' the signatures tha-t it may happen to rabroadcast ln response to the given incoming signature.
No node reacts to any sigrlature arriving with a RRPEAT
35 value greater than some limit provided for in the Sel~healing controller logic.
Within the scope o~ this invention, it is possible to general:lze the concept of the REPEAT field in a number of ways to pursue certain desired properties. ~or instance, the ~EPEAT field could be made some real-valued and/or nonlinear cost function updated by each Tandem node to achieve low rerouting path delays, 5 to avoid certain portions of the n0twork, to selectively use or avoid certain transmission facili-ties, etc.
Another possibility is to let the REPEAT limit be som~
function of time ai~ter the first signature is seen at a tandem node, so that network behaviour tends to find all 10 shortest routings first and then is allowed to probe successfully longer reroutings if needed if the fault is not fully restored within the local vicinity permitted immediately after the failure.
The description thusfar has focussed on the apparatus 16 of the Selfhealing network: Digital Crossconnects, Restoration Signatures, Signature Receiving and TraDs-mitting Circuits and a Selfhealing Controller. It remains now to explain in detail the mechanism (or behaviour rules) of each node in response to various 20 signature events that, when performed simultaneously and independently at each node, causes the network to react wi-th coordinated reroutings of the lost traffic.
To illustrate the detailed mechanis~s through which Selfhealing is effected, l-t will be assumed that a span 25 has been cut between two nodes X and Y in an arbitrary network of nodes and re~erence will be made to the diagrams in ~IGURES q to 14 which correspond to this example case. The span cut is assumed -to have disrupted three worklng circuits, causing Selfhealing to seek 30 three re-routings. The slgnature event-driven sequences at the two end nodes will be considered, initially neglecting how slgna~ures were adjusted and propagated at intervening tande~ nodes. In FIGURES 7 to 14, the following notations are used:
TS.~ield denotes a certain field of the Transmit Signature on a given link:
RS.~ield denotes a certain field of the receive S:ignature on a given link;

TS=~...) or RS=I...) is used to enumerate the entire contents o~ El signature, in the implied order NID, SOURCE, TARG~T, INDEX, REPEAT:
"nul" is used to denote a s:ignature or signature field -that is in -the logically inactive or inapplicable state; and "x" is used to denote a fleld that may be de~ined but its e~act value is not of significance in the current context.
S~NDER-CHOOSER ARBITRATION
SENDER-GHOOSER arbitration pertains to the decision diamond in PIGURE 6 labelled "ROLE ARBITRATION".
I~mediately after the span failure on the transmission 15 facility between X and Y, alarms occur at sites X and Y
and the Sel~healing controllers at those sites are given priority over the normal controller. Node X will read the last valid recelve signatures on the alarmed ports and see tbat (from the NID field) node Y is the node to 20 which connectivity is lost. If not all of the alarmed ports show the ~ame NID then we have two or more simul-taneous span cuts and this can be treated as two simultaneou~ ~aults. Likewise, the Sel~healing controller at node Y is activated and see~ that the 25 problem is to ~ind reroutings to node X.
Nodes X and Y each independently pereorm the SENDER-CHOOSRR arbitration and, Y, having higher ord:lnal rank in the alphabet than X, becomes SENDER and X
becomes C}IOOSRR. These isldependently determined but 30 mutually compatible choioe~ o~ SENDER and CHOOSER roles correspond to the SENDER-CH()OSER arbltration block and the SENDER and CflOOSER states shown schematically in ~ I GURI~ 3 .

S~NDBR PORWARD SIGNATUR~ PLOODING
Node Y then commences the forward flooding wave as schematically illustrated in PIGURE 7. Using the notatlon deflned above, PIGURe 7 illustrates the general manner in wllich a SENDER node initiates the ~orward flooding wave o~ a Selfhealing event. The importan-t things to note are only the RS.NID of the ~ailed working circuits are involved, no other receive slgnatures 5 apply, the flooding involves transmit signatures on spare links only and each logical span departing from the SENDER site is ~looded with signatures up to the minimum of either the number of the circuit restorals needed or the number of spares available on the given 10 span.
Slmple flooding of all spares also works, but it is advantageous in overall network restoration speed to pereorm the slightly more conservative flooding shown in ~IGURE 7 and described above. It can be shown that the 15 expected restoration coverage is not reduced by restra-ining the elood according to the above method.
Each forward ~looding slgnature is given the eollow-ing attributes:
TS.NID = Y (i.e. the name Oe the SENDER node);
TS.SOURCE = Y (i.e. the naMe o~ the SENDER node);
TS.TARGET = X (l.e. the name of the ad~aoent node su~fering the span failure. This is also known as the CHOOSER node ln thls context);
TS.INDEX = i where i is an arbitrary but unique sequence of integer lndex values, no two TS's having the same INDEX. In the example o~
~IGURE 5, elght s:lgnatures are ln:ltiated by node Y, and they are indexed simply as 1 to 8);
TS.REPEAT = 1 all originating ~lood TS's ~et repeat=1.

C~OOSER REVERSE LINKING SEQUENCE ORIGINATION
I~ the details Oe how Tandem nodes react to signa-tures arriving at their si-te are temporarily skipped 35 over, except to say in general that they rebroadcast signatures and increase the repeat counts, then, eventually, in any network with suf~icient spare links, signatures of the above SOURCE and TARGET palr arrive at the node tllat adopted -the CHOOSER state, Node X in this case, in response -to the initial ~ault and subsequent arbltration based on node IU's. When such a 31gnature arrives at -the CHOOSER, FIGURE 8 shows the basic manller 5 in which the CHOOSER responds.
Regardless o~ the RS.NID, if the RS.REPEAT Eield is less than the repeat limit, the RS.TARGET node is equal to the node ID of the CHOOSER and the RS.SOURCE is equal to th`e SENDER node to which connectivity has been lost, 10 viewed by the chooser, then the CHOOSER replies by originating a transmit signature that is the complement of the new receive signature, on the transmit side of the same interface port, with TS.INDEX equal to RS.INDEX
and TS.~ID equal to the node ID oE the CHOOSER and l5 TS.REPEAT = l.
In the example o~ ~IGUR~ 8, RS.INDEX = 2. In this case, the CHOOSER will not subsequently respond to any other signature~, on this SOURCE-TARGET pair, tha-t have an INDEX = 2, as long as the given receive signature 20 persists. (In some instances, however, it is preferable ~or the CHOOSER to recognize another receive signature (RS) with the same INDEX if the other RS.REPEAT field is lower, and in such cases to effectively "move" the responding TS to the preferred inter~ace port.) The CHOOSER node will respond to each signature with correct SOURCE and TARCET Eields as descr:lbed above until the number o~ responding Tran~mit signatures is equal to the number of working links cut by the eault, or until no more spars links are available at the 30 CHOOSER site. Because oL' the action of Tandem nodes, the CHOOSER may see an incoming signaturs on which it has responded, change its RS.REPEAT fleld to a smaller value, with no other change in RS fields. The CHOOSER
continues to respond with the previou~ly emitted TS in 35 this case.

~q~
- 3~ ~
SENDER REVEKSE LINRING SI~NATURR R8COGNITION
Again deferring -trea-tment of the action of the plurality nf Tandem nodes involved in Sel~healing, the bas.lc action of such nodes causes the transmit signa~
5 tures emitted by the CHOOSER above, in response to forward flooding signatures, to be propagated back to the SENDER node. FIGURE 9(a) shows the SENDER node after flooding as in the example o~ FIGURE 7, and FIGURE ~(b) shows a receive signature which eventually 10 returns (as a result o~ a TS at the C~IOOSER and the action Oe Tandem nodes), to the SENDER. It is a property of the Selfhealing method tha~, through the behaviour o~ Tandem nodes, any such signature returning to the SENDER will have the complement SOURCE, TARGET
15 pair and the same INDEX as the original TS issued in forward flooding by the SENDER.
As long as the above conditions are met for the arriving signature at the port on which it arrives, and RS.REPEAT is below the limiting distance criteria, the 20 SENDER will recognize this occurrence as success in an individual path rerouting to the CHOOSER, and the SENDER
will immedlately operate the appropriate crosspoints to substitute the newly ~ound route through this spare port for one Or the por-ts that was involved in the traffic 25 af~ecting ~ailure. as shown in PIGURE 9(c). Although the SENDER node does not know where the restored traefic slgnal will be routed through the network, it is the property o~ this .InveTItion that when the above cir-cumstances occur at the SENDER, it is known with 30 certainly that in ~act whatever the actual routing, the other end of the new path ~ound is at the CHOOSER and the path involves less that Repeat limit minus 1 other DCS nodes operating in concert to achieve -thi~ rerouting (although llkewise ~or them, they have no more knowledge 35 o~ the overall picture than does the SENDER).

~INAL SIGNAL MAPPI~G AND MATRIX CONNECTION AT T}IE
CHOOSER

Soon after the SENDER operat~s the crosspoint as shown in PIGUR~ ~(c~, a live trafEic signal, carrying signal identifier information, as known .in the tel~com-munications industry, arrives at the CHOOSER site, 5 through the above rerouting path. Wh~n this signal arrives at the CHOOSER, the CHOOSER can deduce from the Signal Iden~ifier borne by the carrier signal, which of the possible many ports affected by the span failure, is the correct one to reconnect this s:Lgnal to through the 10 switch matrix. When this is done, the CHOOSER node operates a crosspoint as shown in FIGURE 10, completing the restoration of one traffic carrying signal entity that was disrupted by the ~ault.
The final stage of mapping signal payloads identical-15 ly at each end to the new circuit routings found may notbe obvious but can be handled in several ways. The problem basically is this: If, say, three links are lost in a span failure between X and Y as in FIGVRE 7, and the Selfhealing network responds arbitrarily quickly 20 to provide three path reroutings between nodes X and Y
through the rest of the network, there still remains the further problem of ensuring that both ends, X and Y, use the~e three reroutings with the same mapping of failed traffic signal -to restora-tion path number. The follow-25 .Ing methods can be used to address this last considera-tion.
In the preferred implementatioIl, every traefic carrier slgnal such as DS-3 (44.73~ Mb/s carrying C72 voice circuits) or a SONET STS-1 (49.992 Mb/s carrying 30 700 voice circuits) is transparently encoded with a unique network-wlde slgnal identifier. When this is done, the appearance of Signal ID at the CHOOSER, after the SENDER operates i-ts crosspoint, immediately indi-cates the port to which the CHOOSER should substitute 35 the rerouted signal, because the same Signal ID :is seen on the corresponding interface port as in ~IGURE 10.
This arrangement is the most general and works without ~?~91~9 -- ~o --requiring any prior link numbering convention be~ween DCS nodes.
If all lillks on a span are assigned some ordering number, known -to both DCS nodes on the given ~pan, then 5 the final trai'fic subs~itution sequence can simply be performed in the order of assignment on that span.
However, this requires the administrative coordination of a link numbering scheme for each span in the network and maintenance of the agreed numbering sequence at both 10 end nodes.
In the special case of a single link failure, this whole consideration is of cour~e unnecessary because there is no final mapping ambiguity.

15 TAND~M NOD~ ~HAVIOUR
Reference will now be made to FIGURES 1~ TO 14. A
node which enters into a Selfhealing restoration event in response to a signature appearance on a spare circuit, as opposed to an alarm on a working circuit, 20 enters the Setup-Tandem state in FIGUR~ 6 and is referred to as a "TAND~M NODE" for purpose of descrip-tion.
~ or simplicity of discussion, it will be assumed that all slgnatures involved have SOURCE=Y, TARGET-X or the 25 complement: SOURCE=X, TARGET=Y. Any signa-ture~ arriving at a Tandem node in this example with dif~erent SOURCE, TARGET fields either (a) pertain to another simul-taneously occurring Sel~healing event, in which case they are processed separat~ly, in the context Oe their 30 respective faults, or they are simply ignored, or (b) the signature i~ spur.lous or erroneous and i~
ignored. Any signature arriving with RS.REPEAT equal to or greater -than the limiting criteria will be ignored and this case will not be discussed further. In 35 addltlon, any recelve slgnature that is said to be "repeatable", will cause one or a number of new transmit slgnatures which will be given a repea-t count in TS.REPEAT that is the RS.REPEAT of the corresponding RS

incremented by one. These rules of repeat limiting and repea~ incrementing and SOURCE/TARGF.T L'ield oons.istency apply throughout the remaining aspects o~ the method at the TANDEM node and wlll not be ~pecl~ically 5 re-iterated.
Likewise, in all case~, the RS.NID ~ield is used simply to deduce span associations amongst independent links (i.e. all links at a DCS e~hibiting RS.NID=K are known to be on the same span or equivalently have the 10 same immediately physically adjacent DCS node destina-tion) and TS.NID i~ always set to the node ID of the given Tandem node. There~ore, NID fields will also not be explicitly discussed again, although use o~ RS.NID
~ield is implicit whenever behaviour is de~cribed in 15 terms of span oriented kno~ledge. The above considera-tions simplify the description o~ Tandem node behaviour into terms dependent only on the use of the INDEX field and the ASSOC-PORT registers on the DCS inter~ace parts.
Be~ore describing the TANDEM node behaviour, it will 20 be noted that the same node is equally able to ac-t as SENDER or CHOOSER in respect of a span fault ad~acent to itself. When it acts as TANDEM node, it is not directly involved in a fault but it is acting on behal~ Oe the two nodes -that are involved directly in any given span 25 ~ault. Although the actions o~ only one TANDEM node are described here lt is implioit that the same behaviour oacurs, according to the same principles, simul-taneously at any number o~ other TANDEM nodes to e~ect sel~-healing.
BA8IC TANDEM NOD~ SIGNATUR8 REBROADCAST
FIGURE 11 shows the basic slgnature rebroadcasting behaviour o~ a TANDEM node. When a signature is received, its RS.INDEX ~ield is used to Eirst check 35 whether or not the node ls already broadcasting any TS
signatures with this INDEX. I~ not, then the rebroad-cast pattern shown in FIGURE 11 i~ established. In this pattern o~ limited signature rebroadcast, one copy o~

the .incoming signature is written to one idle TS
register in the set ot` spare l:inks leaving -the node site on each logical span (i.e. one copy of the recelve signature w.ith that speciic index ls æent to each 5 adjacent node reachable from the TANDEM node site). The REPEAT field is incremented and the ASSOC-PORT register of the Transmitting Signature circuit on the selected ports are written with the identification of the por~ of the incoming signature that caused these transmit 10 signatures.

TANDEM NODE RECOGNITION OF A BETTER PRECURSOR
As the dynamic process of Selfhealing proceeds and the volume of signatures impinging on a given TANDEM
15 node increases through forward flooding, it sometimes occurs that a new signature may appear at a certain port and that signature i9 a "better" precursor for some existing transmit s:lgnatures than the Receive Signature that originally caused those transmit signature~. In 20 this case, the be-tter precursor signature will take over responsibility for the existing transmit signatures ~or which it is a bet-ter precursor. This notion of better precursor and the takeover of existing TS's by associa-tion with the better RS port is part Oe what gives 25 Sel~healing, at the network level, the property of selecting m:ln.tmum length routes.
PIGURE 12(a) shows three TS's associated (by the dashed llnes) with an RS having INDE~ 2 and REPEAT r.
In FIGURE 12(b), an otherwise equivalent slgnature 30 arrives at a d:l~ferent port w:lth a lower REPEAT f:leld.
In ~IGURE 12(c), the TANDEM node reacts by altering the ASSOC-P()RT registers of the relevant transmi-t s:lgnatures to reflect association now with the bet-ter signature at the new port. In addition, the 'rS.REPEAT fields are 35 ad~usted to reflect association to a precursor signature with lower repeat count. It wi:ll be noted -that no crosspoints have been operated yet at the TANDEM node in respect of the current Selfhealing action. The associa-.5~3 - ~3 -tion between TS and RS signatures at the TAND~M node i9 only a step in the dynamic progression to a sltua-t:lon where the TAND~ node may recogni~e that a certain crosspoint can be opera-ted to help satisey a path cons-truction for a certain IND~X number.

TANDE~ NODE RECOGNITION 0~ SIGNATURE COMPLE~E~T
As a result o~ CHOOSER node Transmi-t signature initiation in response to forward flooding signa-ture 10 arrival, a TANDEM node adjacent to the CHOOSER node will be the first to see and recognize a "complement signa-ture pair". In FIBURe 13(a), a basic TANDEM node rebroadcasting pattern is shown on a given INDEX. As a result of being next to the CHOOSER or as a result Oe 15 the action of other TANDEM nodes performing -the same behaviour as now described for the subject node, it is possible eor the subject node to receive a new signature which it recognizes as creating a complement signature pair on the port where the new signature has arrived.
20 The complement signature condition is de~ined as RS.INDEX = TS.IND~X, RS.SOURC~ = TS.TARG~T and RS.TARGET = TS.SOURCE. In ~IGURE 13(b), this is intended to be the case at port 132.
~hen the above complement signature situation arises, 26 the TAWDEM node in this invention then does all of the following immediately, these actions beinLr shown schematically ln ~IGURE 13~c): A matr:lx crosspoint palr is operated connecting (bi-directionally) the port where the complement-creating receive signature was received 30 to the port identif:led ln the ASSOC-PORT ele:ld Oe the TS
half Oe tlle port o~ the new signature; all other transmi-t signatures associated with the precursor receive signature o~ the current complement ~ignnture port, are suspended and the association reg:ls-ter 35 contents are nulled; and a transmit signature is now applied to the transmit half of the port where the precursor receive signature resides. The transmit signature is identical to the receive signature that - 4~ -caused the complement to occur, except that its ~EPEAT
field is incremented.
In summary, when a TANDEM node receives a signatllre that causes a complement, it prunes off the unneeded 5 branches of the broadcast tree on the given index at its site, extends the path of the complement signature in the one direc-tion where it continues -to produce a complement signature, and operates crosspoin~ connec-tions between the two uniquely identifiable ports 10 involved. In this way, it can be appreciated, at a larger network scale, how the forward flooding process sets up a complex tree of possible routings from all points back to the S~NDER, and then how the reverse linking signatures selectively emitted by the C~IOOSER
15 cause the actualization of certain of the potential paths, pruning unneeded branches of the tree and reinrorcing others by operating crosspoints and selec-tively propagating the complement condition through TANDEM after TANDEM until the complement signature 20 arrives at the SENDER and the SENDER Reacts as previous-ly described.
Although not shown in FIGVRE 13(c), when a complement event causes the elimination of certain TS originatlons, the Selfhealing logic -then re-considers all exist;lng ~S
25 fields that are not complemented as possible precursors for new TS signatures to be applled to the newly freed TS registers. Therefore, according to the behaviour given for bas:lc signature rebroadcast and bet-ter precursor takeover, any TS port freed as a result of a 30 complement on a certain INDEX, may be immediately pressed into use on another persisting INDEX that has not been satisfied.
It is within the scope of this invention that the above mechanism (which is of great complexity when 35 treated fully at a network level, but relatively s.imple for any one node in isolation) will always, after signature distribution through SENVER floodlng and TANDEM rebroadcast, as already described, succeed in - ~5 -propagating the complemellt conditlon back to the SENDRR
node, through a minimum number ol' TAND~M nodes, once the CHOOSER emits an appropriate reverse-linklng sigllature.

~ISAPPEAR~CE 0~ A SIGNATURE AT A TANDEM NODE
One remaining event that can occur at a TANDEM node is the disappearance o~ signatures or the complete replacement of one signature by another signature with different index count and/or other fields a8 well. The 10 signature replacement case is handled by the TANDEM node logically the same as a sequence of signature disap-pearance and new signature appearance. The response to new signatures has been descrlbed, so with the Eollowlng description o~ response to signature disappearance, the 15 description of isolated signature reaction behaviors will be complete.
Signature disappearance can occur either as the direct or the propagated e~fect Oe TS removal at other nodes such as the SENDER, as the required circuit 20 reroutings are obtained, or such as at a TANDEM node that has recognized a signature complement and stemmed off certain TS originations.
The reaction -to signature disappearance is i'a:lrly simple: when an incoming signature disappears, any 25 transmit signature associated to it (FIGUR~ 12(a)) are simply disconnected ~rom any association and the TS
register contents are nulled out (~IGURE 12(b)).
Immediately therea~ter, any other present incoming signature on another lndex whose rebroadcast pattern 30 could be use~ully augmented, is permitted -to takeover assoc:lation responsibllity Eor some or all of those transmit signature registers, as in FIGURES 11 or 12(c), and have the appropriate TS register contents applled to those ports. In general, therefore, when any 35 receive signature disappears, any transmlt signature ports that were assoclated with lt ~orm a pool of new ldle TS registers to be considered ~or TS signature LR~:~

- ~6 -application in general by any o~ the preceding rules g:1ven.

6 APPLIC~TION TO SELF-PROVISIONING NETWORKS
The mechanism described above can also be employed or automatic provisionin~ o~ new circuit routes in a telecommunications network to deal with circumstances Or unanticipated -traffic load between certain nodes, rather lO than loss of existing facilities throu~h a failure.
This is done by including within the implementation an optional means whereby two nodes anywhere in the network, between which it is desired to provision additional circuit routes, are placed directly into the 15 SENDER and CHOOSER states with regard to an artificial fault between the selected nodes. The artificial fault information includes the number of circuit routlngs tha-t are being sought. Every-thing else functions as des-cribed above.
Application o~ this invention in a network such as those in ~IGURES l and 2 makes it possible for these networks to dynamically and continuously adap-t the provisioning of physical circuit routes within the network to the actual time-varying point-to-point 25 tra~ic loads within the network that are measured statistically by -the lower-level call-by-call traf~ic switches. r~, for instance, the voice-call levcl switch in Toronto were to recognize excess probability of blocking on the circuit group routed to Edmonton 30 (perhaps because of an unpredictable event caus:Lng focused overload, such as a tornado), the Toronto and Rdmonton DCS machines can be st:1mulated to perform an artificial Selfhealing event, thereby obtaining addi-tional transmiss:10n routings -through spare capacity 35 available elsewhere in the network in a time in the order of seconds. When these extra -transmission facilities are configured, the view ~rom the telephone traffic switches (call-by-call swi-tches) is thereafter - ~7 -.identical e~cept that the logical trunk group sizes on the subjec-t rou-te are suddenly larger in size, thereby adequately reducing the probability of call-blocking on that relat:lon (Edmonton-Toronto, for example). In 5 today's network it can take ~rom days to years, depend-ing on equipment, operating methods and the source of the overload, to recognize trafEic loading conditions that are not according to engineering foreca~ts and to react with the facility changes required to increase 10 physical capacity between affected nodes. It will be noted that this is dif~erent from dynamic routing of traffic within fixed size trunk groups, which is known to those slcilled in the art of telecommunica-tions.

APP~ND I X A

SI~L~H13ALI~G NleTWORK DCS
CONTROL PROTOCOL

3~ ~J L?~ ~
PIOC~l~ 3N t~fi~ ; ~. Crover, B. Venableg Sept ~, 1981 CO~S~ nports : DCS si~el;
~a~repeats - allo~ed reacb for Selfbealingl;
~a~gpansi~e : lar~est span at tbis nodel agna~e : ~sslgn ne~or~ ~ide 10 of t~is nodel Setup ti~eout : Sel~e~ling ti~e out tbresbold 2 secl ~tce Interv~l : optionsl re-erecution inter~al a6ter ~etupl S~P8 nodeid : (enu3erated list o~ vAlid node nà~es plus a nul vsluel;
current state~ : (Not~al, Setup tandea, Sender, Cbooger~;
portids~- I..noorts~ nul;
signalids : I Definltion accordi~d to net~or~-~ide rD scbe~e, plus t nul vnlue BSre~ister : B8coaD
NID :nodeid sourcaDod,e :oode}d;
targetDoae :nodeld;
iDde~count :O. nports, nul repeltcount :O..aasrepelts~. nul;
alar~, spate, ~Sdelta, 9iglDdelta :boolean;
SignallD :slgntlids 58registet : BBCOBD
~ ID :nodeid sourcenode :nodeid;
tartetnode :nodeld;
indercount :O..npotts, nul;
repeatcount :O..oalrepests~l, nul;
asaoc port :portids;
Si~naIID :signali~s 8~u;
interrupts : I v~lid interrupt vectors in tbis 051;
~lR B9: lBBAY ll..-~portsl OP E9register ;
TS: lBBl~ [I..elrpottsl OF TSregister ;
Dortnaee~ lnterrupt_port : portlds ~ftected_Port~ : arra~ll..ua~spsnsirel of portid~;
il lostccts, restcct~ : Integer;
S~_interrup~_vector: Interrupts;
2TBBPlL
ptocedure Load tlaer(lnter~all;
proce~ure 8tar~ tiner;
proceaure 9top_~luer;
nas~ Si~lDde~alport:portldsl;
enabre ~ Duel~a(Dor~:Dortlus);
Drocedure Operlte 2~a~ ~ro~spolnt(portl,port2: portids~;
Procedure ~ait Suspend~l9N interrupt vec~or:interrupts, interrupt Dort:portids);
[6uspends S~-tas~ in OS . ~ill return on 4ert Instlnce o~ 0 Interrupt vector : S~ interrupt vectot, ~Itb ~D of port tb~t generated~interruptl ~
B8CI~ ~ one tine installetion initinlirationsl for i :: I to ~rspansi~e do lftected_Portslil:: nul;
~OB i :: I to nportà DO
B9 i .sourcQ :: DUI;
B9 i .tat6et :: DUI;
~9 I .indercount :: nul ;
89 i .repeatcoun~:: DUl;
~9 i .P9delt~:: rnlse;
~9 i .9iglDdçlt~::t~lse;
n~ ~ -8iglDdelta~
T9 i .source :: nul ;
T9 i .tsr~et :: nul ;
T8 i .iDde~oount :: nu~ ;
T8 ' .repe~tcount:: nul ;
~9 i .~ID :: n~n~re;

s~

~S alsre is bardllare intitiallized ~ conkolledl tS¦l] ~ssoc port ~nd l`S/BSIl~ Si~nal ID6 are l/ritten in noroal operation b~ t8e crosspoint operator proceis 3nd ntlallzed pr~or to self-beallng accor-din~ to tbe norual configuration ot the cros6connectl 1~$111 NID arriYes on BI direc~lon signal - selD-intialiainll orl ~PBlt ~ islinite repellt loop repre~entiog interr8pt-driren e~ecutioDl ll~it_Suipend l91_irterrupt_7eGtor, interrupt_portl;
Dortna e: Interrupt port;
nor-al NorualNode;
Sender Setup 9ender;
C~ooser 9etup-C8006er BRD ICA9B~getup_t~nùe~ Setupl~nde~Nole gS~Iortn~nel .aSdelta::fsl~e;
U~ L ful~e;
P~OCI~DU~X C~ec's lad_Ti~rl Var portno portidsl;
BBCIN
IP eSlportnol sDare gSIportDo~ tns~etnode~)~SIportnol 30urceno e) 0~ K~ portno Indescount~S9 portno Indescount 0~ 19 portno sourcenodel)TS portno t~r~etnode ¦ tH8N
tS portno t~r~et~ode :nul;
TS portno sourcenode :nul;
tS portno indescount :nul tS portno repeatcoùnt DUI;
9RD t7 portRo ~saoc port :nul;
BRD; IP~C~ lULB C~ ct_And_rid~l IIOCIDU I 8ilCOP~Plood (portD~e~;
BBGIN
Poa iode~ :l JO s,ports DO
IP l~i4desl sparel AND la91in,desl 11 D~)llSlportnlleel RID
8ABD ~O~Itegl~indesulsilelBdlN Atnt~ RD~H t91indesl n~soc port nul~
tS ,indes repentcount :R9¦portnnuel repeatcount t 1;
t9 !ndes tartetnode :~S portntoe tsrletnode;
tS Indes sourcenode :g9 portna-e sourcenode;
~9 indes LndescouDt :g9 Dortna~e indescount;
tS indes ~ssoc_port :Do tna~e~
BN Nodlelp AttNodo t Ig91inde-l ~IDl;
B~D', ¦IOgl 8ND; Procedure Si~COPrrlood/
ZOCBDUII gorualllode; Inor-alnode ia called ~Iben tbe node b~s no previoua slg co-in~ into itl rPB events:lsender~l~r~, cboosernlnr~, repe~table slg reoieved, noDrepeatable_all_reoelved, ~purlous~' ~
VA~ IttRode 9B~ or Dodeid ' Inde~, Indes ~tnllp inte~er;
evtnt events, PIOCIDU~B 8ill01ICllood llnde- stnllpl;
B 8t Ct I N d I I
FOII indes I TO nports DO
BBCIR

9~
IF 'indesl spare) AND NO~ ~B91iDdeY]~al~rD I
~D ~B9~ndesl~sourcenode: nul) I~CNNOT tBSlinde~l IIID IN lttllode) SHBII
SS indes repeatcount: 1;
~S indes tarl~etnode: ~SIportnanel~NlD;
~g indes soutcenode: s~naue SS indes indescoun : indes s~--lt NDIt'', lttllode ~ BSlinde~r R[D , IND; ~ FO~ I
8ND; IProcedure Si~O~lOfloodl PBOCIDUII~ 3eil~ lffectc~ Port3 T~ble;
IScan6 all DCS pQrts o find ~n~ nei alarc3 on uorsing ports count,s tbee llostccts~, stores tbe Dortnu3ber6 of t3ç ports t~e6e fllile t~cili~ies uere connected to tbrou3ù t~e DntrlI and reset6 tlle interrupt li~e ~9delta lor all ports ~t nas loUn al~r~le~ I
811Cl~
lottcct~ ~ :0 80B iDdes :: I SO Dporti DO
If a$1indesl elare aND NOT IRSIindesi spare) AIID IQ91in~esl BSdelta ~RBN
8~CIII
lostccta: lo~tccts ~1;
,r ec~ed portsllostcctal: r91indesl.a6soc_port ~S Indesr~Rgdelta: tal6e BRD;
BllDjlPBOC3DUB1 9uilt_Affected_~orts_raole; I
B8CIII IPBOCRD~B2 ioru~lgoae;l If Itst decide ubicb t7pe of ail ia beinl recievedl lB 1(~81portn~eel.tar~etDode null AIID lR81~ortnnee] sourcenade:null AllO NO~ Bglpo~tna-el alaral ~N8N event :spurious Bbg38C~ll 119lportna-el allr~ ~'dBh Build Alfected Pott~ ~uble;
ae~ Ja ~e)t81porta~e]-NlDI ~L88N e~eantt.-c~eOO8eerl;ar~;

81,93 1~ ~8lportna~el repehcouat(~usrepeats ~NB~
eveot :rePeatable i recieved ~L9B event :aoarepea hle_sit_recieved;
lao~ proceat tbe ~ accordinl to its trpe C191 eveat 0 ~uriolla~ tub I
e-derul~r~
BON iade~ ~tat :1 to la~tccts DO 9i~0~1Cflood (indes ttnopl;
iDdes st~op : ~ ~
Bor iades : I to nports do it ~9 indeY aourcenode ~) null nnd S9 i~de~ ~ndescount ~) au be notll l iades st~up : indes stnap ~1;
S8din~es].lnde~count~:: inde~_~ts~p, tnte :~eader;
9tDrl tilet;
BllD;7seader~1ar d8gol N~
atnte :tetu~ restare;
9t~rt_5iaer,~

- ~ 2 -BND; Ic~006eralar~) reDent~ble sig recie~ed BBCIN
lttnode: I ;
SigCOPYPlood portnasel; ;
state :setup tandei; ~;
BND; IreQeatabre sit_recievedl ooorepe~table sig recie~ed ~BGIR
st~te: noroal BBNB j oonrepeatab~e_sig_recieYed 8ND ! PllOC~DUeB ~lorl~ulllode PBObB ~U 9eto~_9eDder;
~7PB çvents:ltiaer int, Spurious, Beturn si8);
VlB Inde~ IDtegeF;
event:eveots;
BBCIN 19ender 1 IP ltiserl tRBN event :ti~er int aLsB IP NOt 1891Portna-el B8~eltal tHBN eveDt :Spurious BLSB IF ~B81Dor~Daoe] tataetnode ~rna~el ~R8N event :Beturn sig;
BLSR event :Spurious;
C19B e~e~t OP
BB~IN
POB inde~ :l SO DpOrts DO Cllec~ aod tid~(inde~) state :9eDder;
Lt-D-ti er( t i t 1¦
BuD;StlatirFeriienrt;t eporiou~; Istubl BRCI N
ll~t confiro tbnt tbe rçturnin~ si~ is a conolecent IE so, operate crosspolnt to subsltute raEEic toruarding trnEEic (and 5i61DI to tbe Cbooser 1 IB IBalportnaael sourcen~e: ~9 portnaoel tar8etnode~
D8ulB8bportna~el inde~count J91Portna~el~indescount¦
restccts :restccts~
ODer-te Ctoss Poirt ~tected Dort~lrestccts] portotoe~;
1 t91porFDa-el.9i~1g :: B81AE~ected Dortslres~cctsll.91~1D;I
I Eo uardint ot 8ItlD i8 i-plicit urtb operation oE the cros8pointl B lolNtccts: resteot3~ ~UBN
P B inde~ :l tO nports DO Cbec~ ~nd tid~linde~
state: 9ender; ~ ~
LBoND~~tlljelr(atce_intervall;

BL9R state :9ender;
BNq;
BND Ci9~ ~
BNDj ~tPBoC~DUB~ getup_Sendtr;
PIOCBD~B8 8etcp Cbooeor;
~PB eivdots-lotl~for~lttd_sil,sll~id_arrive, 9puriou2 1;
inde~ satis ied boolelo;
eventrevents;
8BCIN 19etu~ Cbooaerl Icbeck~to eee ~Ibat t-oe of si~ is co-inl inl IP B91Dortna~el tartetnode~)a naie ~RBII event: 9purious BL9B IP (a9 portn~llel B8delta ~dBN event: ne~ forusrd sig BL9B IP IB9 portDIlnel 8il~1Dde ta~ ~RBII event :8rglD arrrve -- ~ 3 -~

BLSB event : Spurious ;
c~98ou grtcgFs the sig according to its tgpel purious 9BGIU
Cbec~ lnd Sid~lportnaDe);
BgsD,te: C~Do~er;

De~ for~ard sit -~BCIU
IF restccts~lo6tccts TNBN
BBCIN
,indeI_satisfied::false;
Indes :0;

index -indes tl ;
~5~N in3eiiBat~y8cofuined -trRuepottnalel~indescount UNT11 Indev:nports~ OB indeI satisfied IF NOT index sttisfled THBN
BBGIU
~5 portna2e tatgetnode :RS portnane sourcenode;
T$ portnnae ,sourcenode :~5 portnaoe ,targetnode;
T9 portna~e ,inde~count :18 portnaae indercount;
T5 portnane repe~tcount :l T9 portn,aae assoc_port -portnaae;
Bn ole Sl~l deltalportnanel~
~enable tbe si~nal ID i~nter upt on tbis por~ go Chooser uill be invo~ed uben tbe Sender ODer~tes crosspoint oruarding tralflc,tsotjl centopettalte our crosspoint to ~nit bac~ tbe corresponding si~ID
BUD~ IIFl stn~e :Cbooser;
luil8NstlYf in Cdooielr even if restccts : lostccts until final ai8DallD Dappin8 i8 co~pletedl 9i~id k ri-o jsignallD cbange caused the interrupt Cbooser uill operate BBGIR a crosspoiDt terel iDdeI::O;

inder :inde~
UH~IL E9~affec~ed ~Dorts[indetll,siglD:B91~ortnanel siglD;
Oeerate zlla~ CrosaPDiDt[a~rected_pottslinaexl~portnane);
llfecte~ por~s[indesl :nul;
BND 13igrD_Ar ldieveltnlportniee~
aND; IC1881 BHD; IPeocsov~ getup_Cboosètl PDOC5D~gP ~etnpT~Dde~H,ode;
T~P8 events:Jno si~, spurious si8 vanisbed, ne~ coepleaent 8il, repeatnule_sig, nonrepea~nble~sit~; ~
VA8 AttUode 8R~ of nodeid;
indeI Inteter involqed, aifferentsig boolean;
event events;
FU~CT101 3etterHi~(Dortnrnel trLd ~ st~ ~ ~ t ~ i~ ~ sii~ on portniaae is tbeibest of its ~ind co~ing in fot n ~iven indescount indes Og RBPRiATd icdapalre current R~sig on portnaee against all otber portsl IF E91inde~ sp3re A~D II~RS~Indesl indexcount : RSIportnnue] indexcount~
AN IBSIIDdesl sourcenode : RSIportna~el sourcenode~

- ~ 4 ~ ,9 l~D indeY~)portna~el lUD BSlindeYI.targetnode: RS[portnanel.targetnode~
lND IBSIinde~l.repeatcount~: ESIportnalel.Sepeatcountl) OR l(tS Indesl.indeYcount:R91indeYI.lndeYcountl ~ND TS indey .sourcenode.BS indeY ,targetnode lND tS indeY .tar,eet'node:RS indeY .sourcenode )) tHBN ~etterSit :t ue;
UN~ID linde~ nportsl OB BetterSig;
R~D; 18etter-sig-searct1 B~OCgDUR8 Cnncel ~ Si~a Fro~ ~port portna-e)i tcanc~6-t~gl~ar~nFFil;~g~tlro~r"one-"R~6l2""~at-"bag diaappeared or been over~ritten. Uses asociated port fiel'd of TS register'to quic~lg indentif~ tS6i~S ttat ~ere caused b~ tbe BS at portna~e. 'tbe' acan of all ports ia used as an opportunit~
to see if t~elt re-ains an~ sitnature activit~ at tbls node. If not, involved ~ o false and cause a return to Noraal state I
involved::false;
POR inde1::1 SO nports DO

18 ~91inde-].spare lND l5S~indeYI.assoc port:port AND llnde~ () portnaoe¦TRB~
B8CI~
TS indes .sourcenode::nul;
TS indet .targetnode::nul TS inde~ .re~eatcount: nul;
T9 indel .inaescount::nul;
END,tlS pnlde~ .assoc oort::nul;
IP ¦It IndeYl.sourcenode()null BUD~ IFOR BSIinderl.sourcenode()nul)) TR8N involved::true;
3YD; ICanoelr~S g61 Procedore U date Sendine to aet lioderco~otl;
t-olrird~!7 FF ~ f-n-~te-I7~ r-~ict~-tb-i~-b-o-d~e-i-6~ currentl~ sending 6i6nature6 on a 8iven inde~count. rbis set used to ~annge slgn~ture rebroadcastinJ I

lttnode :: 1' FOB inde~:: TO nports DO
BBGIN
IF IT91indeYI.sourcenode R91pottnanel.sourcenode) AND (TSIinde~l.tartetnode:B9[portnaael.targetnode lND lJ9llnde~l~inde~count:R9lportnaael~indeycount tHEN
DBGIU
AttNode :: AttNode,~ IR91inde~l.NIDI;
ta~e tois oppottunitr to ta~e over outgoin~ signatures for ~bicb I aa actuall~ a better predecessorl IF l(T91inde~l.repeatcount)RSIportnaoel.repeatcount~l B GIHND IRSIindeYl.NID()RSIportnaael.NIDII ~REN
t8 inde~ .repeatcount::RSlportnauel.repeatcount~l T9 iDde~ .aas~c port::por~na~e;
END;
END
BND; (FOB
END; lUPdate_ 'endin8_to_setl BBCIN IBetu TaadeoHodel 't~-FFC~Fl~F ~ ~tlrFZr ~6'a port bas ctanged and tbe SN tas~ RaB in Setup-rsodea 6tate.
re arrive here ~ith portDaee set equal to the port ~bere this cbange oocurred.l ,Y8PEl~ I rhe ti~best le~el here is a loop tt~t causea Droces6iDg of BYB~7 ports ~S
iD a (aodulo Dport61 sequence starting fro- tbe current PORrUAHB : ieJ the port on ~hicb tte ~S deltt that caused an interrupt ~a6 seen.l P~fl~I~C Y~,,.,,,l nvolved::true~
IE RSIportn~ael.RSdelt~ ~HgU ditîerentsit~:true EL5E differentsia.:false;
18 IBSlDortna-el.t~rtetnode:null AND ¦R9[portnnael.sourcenode:nul¦
rRgN DgGIN IB dlfferentsi~ ~RBN e~ent:: alg vanis5 BLSBIII 8L98 event :: no~sig;
DECI~ Ichech for a coapleaent 8ig inconingl IB (BS portnane] Eourcenode: TSlportna~el targetnode lND RS portnsue targetnode: TS portnane sourcenode AND RS eortnaDe ~indeYcount: rs portnane indeYcount THBN HB IN Isig n portnaDe is a conpleee t I
IS dlflerentsig ~HBN event :neu CO~pleDent 6ig BbSB event :nonrepeatable_sig;
BLSal21 8BCIN
IB di~erentsil, ~HBN Cancel ~I Sigs froD lportnane,portnaRe~;
IP BetterSIl TH8N event :nonreFeata~le giL
Bl,SB IF IRS[Portnaee] repeatcount~laxrepeats' ~HBN
ever,t::repeatable S18 ~ND BIBLsB2vlent :nonrepeatab1e sig;
BND IBISBII
~ 8~1) OB BVB~,lr PII~SINC r8BB
C158 e~ent 0 sputious ; ~stub~
9i~ ~nia~ BBGIN
~ C~ncel ~ Sigs Fron Iportnaee,portnaoel;
lF NO~~involve~ TNBN state :noraal;
8ND; ISig vaDIsBl co~pleDeDt sil Ipropa8ste tbe inco-ing ~ig in the ditection of its co~plenentl Cancel S~ Sigs Brol (portnsne portnaue~;
inder ~rSrportnaoe] as60c por~;
Cancel ~ Sigs Broa (inde~, portnaue~;
~S inde~ targetnode :BS portnaael targetnode;
tS Inde~ sourcenode :BS oortnaseJ~sourcenode~
~S inde~ repestcount :B Iportna-el repeatcount O I;
SS indes inae~count :BSlportnane] ind ~count;
~et~,e_ et ~ a7~7 ~ dPer-tr~66~gD~ $eeY ~statel BND; Ico~pleaent sigl repo~table ~i~
Isend tte slt doun all clear outgoing lin~s ubicb bave not alread)~ been used, not tuo to lnJ oDe oode dnd none to tbe Dode tbe ~S is frool Update Sendin~ to Set;
8ltcopnlood ~porFns~e);
8ND; Irepest~ble_slg receivedl nrlC15BIble_~il_rocei~od ; atate : setup_tande~
gortnaoe : succ(~ortna~e~ odulo urap 5UCC fUllCtiOD
N~IL port~aoo : l~terrup por ;
BND; ISetuprande~Nodel

Claims (78)

1. A method of restoring communications between a pair of nodes in a network having an arbitrary number of nodes and an arbitrary number of spans interconnecting said nodes, each said span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits between nodes capable of, but not designated for, transmitting actual communications traffic, said method comprising the steps of:
a) establishing one or more independent communication paths between said pair of nodes through a series of spare circuits of spans interconnecting said pair of nodes and other interconnected nodes in said network; and b) redirecting communications traffic intended for one or more failed spans interconnecting said pair of nodes through one or more of said paths.

2. A method as defined in Claim 1, wherein said establishing step comprises the steps of:
a) repeatedly transmitting restoration signals along logical spans of spare circuits departing said one node;
b) repeatedly retransmitting said restoration signals from said other interconnected nodes along logical spans of spare circuits departing said intercon-nected nodes; and c) upon receipt of a restoration signal by said other node, transmitting a complement restoration signal from said other node along a path consisting of the same spare circuits along which said restora-tion signal was communicated to said other node.

3. A method as defined in Claim 2, including the step of identifying spare circuits at said one node and allocating a distinct restoration signal to each identified spare circuit.

4. A method as defined in Claim 2, wherein said repeatedly transmitting step includes repeatedly transmitting said signals along each logical span of spare circuits departing said one node up to the minimum of either the number of circuit restorations required or the number of spare circuits available to said one node.

5. A method as defined in Claim 2, further including the step of continuing to respond, at said other node of said pair of nodes, to restoration signals until the number complement restoration signals equals the number of failed working circuits or until no more spare circuits are available at said other node of said pair of nodes.

6. A method as defined in Claim 2, each said restora-tion signal and its complement restoration signal including an attribute representative of a particular one of said paths.

7. A method as defined in Claim 2, each said signal including an attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path.

8. A method as defined in Claim 2, each said restora-tion signal and its complement restoration signal including a first attribute representative of a parti-cular one of said paths and a second attribute represen-tative of the number of spans and/or linking nodes in each said path at any given point along said path.

9. A method as defined in Claim 8, each said restora-tion signal and its complement restoration signal including third and fourth attributes identifying said one and said other of said pair of nodes.

10. A method as defined in Claim 9, each said restora-tion signal and its complement restoration signal including a fifth attribute representative of a network-wide identification of the node originating said signals.

11. A method as defined in Claim 2, further including the step of:
a) designating one of said nodes as a SENDER node from which said restoration signals will originate and the other of said pair of nodes as a CHOOSER
from which said complement restoration signals will originate.

12. A method as defined in Claim 11, wherein said designating step being on the basis of an ordinal test.

13. A method of restoring communications between a pair of nodes in a network having an arbitrary number of nodes and an arbitrary number of spans interconnecting said nodes, each said span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits between nodes capable of, but not designated for, transmitting actual communications traffic, said method comprising the steps of:
a) designating one of said nodes of said pair of nodes as a SENDER node and the other of said pair of nodes as a CHOOSER node;
b) establishing one or more independent communication paths between said SENDER node and said CHOOSER
node through a series of spare circuits of spans interconnecting said SENDER and CHOOSER nodes with other nodes in said network, including the steps of:
i. repeatedly transmitting restoration signals along each logical span of spare circuits departing said SENDER node up to the minimum of either the number of circuit restorations required or the number of spare circuits available to said SENDER node;
ii. repeatedly retransmitting said restoration signals from said other nodes along each logical span of spare circuits departing said other nodes;
iii. upon receipt of a restoration signal by said other node, transmitting a complement restoration signal from said CHOOSER node along the logical span of the spare circuit upon which said restoration signal was received;
iv. retransmitting said complement restoration signals from said other nodes along the logical span of the spare circuits on which said restoration was received and connecting the port upon which said restoration signal was received with the port upon which said complement restoration signal was received whereby to permit communications traffic therebetween;
each said restoration signal and its complement restoration signal including a first attribute representative of a network-wide identification of the node originating said signals, a second attribute identifying either said SENDER node or said CHOOSER node as a SOURCE node, a third attribute identifying either said CHOOSER node or said SENDER node as a TARGET node, a fourth attribute representative of a particular one of said paths, and a fifth, incrementable attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path; and c) redirecting, through one or more of said paths, communications traffic received at input ports of each of said SENDER and CHOOSER nodes and intended for transmission through one or more failed spans interconnecting said SENDER and CHOOSER nodes.

14. A method as defined in Claim 13, including the step of identifying spare circuits at said one node and allocating a distinct restoration signal to each identified spare circuit.

15. A method as defined in claim 14, wherein said restoration signals transmitted along said spare circuits departing said SENDER node differ from one another only by the value of said fourth attribute.

16. A method as defined in claim 14, wherein said retransmitting steps including the step of incrementing said fifth attribute of received signals prior to retransmitting said signals.

17. A method as defined in claim 14, further including the step of rejecting at either said other nodes or said CHOOSER node any received signal whose fifth attribute exceeds a predetermined value.

18. A method as defined in claim 14, storing at each other node the first received restoration signal with a particular fourth attribute and rejecting any other received restoration signals having said particulars fourth attribute.

19. A method as defined in claim 14, storing at each other node the first received restoration with a particular fourth attribute, rejecting any other received restoration having the same first to fourth attributes but a larger fifth attribute and replacing said stored signal with any new restoration signal having the same first to fourth attributes but a smaller fifth attribute.

20. A method of restoring communications between a pair of nodes, between which communications have been broken, in a network having an arbitrary number of nodes and an arbitrary number of spans interconnecting said nodes, each said span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits between nodes capable of but, not designated for, the transmission of actual communi-cations traffic, said method comprising the steps of:
a) upon occurrence of a communication failure in one or more circuits in one or more spans between said pair of nodes, designating one of said pair of nodes a SENDER node and the other of said pair of nodes a CHOOSER node;
b) repeatedly transmitting restoration signals on the outgoing transmission link of one or more spare circuits departing from said SENDER node, each said restoration signal including the identity of said SENDER node, the identity of said CHOOSER
node, an INDEX value representative of the identity of a complete, independent route between said SENDER and said CHOOSER nodes starting at the spare circuit departing said SENDER node along which said signal is transmitted, and a predeter-mined initial REPEAT value;
c) upon receipt of one of said restoration signals at a TANDEM node:
i. determining whether said TANDEM node is already transmitting a restoration signal having the same INDEX value as the just received restoration signal;
ii. ignoring any restoration signal having the same INDEX value as that of a signal said node is already transmitting;
iii. if said node is not already transmitting a restoration signal having the same INDEX
value, producing a modified restoration signal by incrementing said repeat value of said restoration signal and repeatedly transmitting said modified restoration signal along spare links departing said TANDEM node;
d) upon receipt of a modified restoration signal at said CHOOSER node, repeatedly transmitting a complement restoration signal from said CHOOSER
node back through the same spare circuit along which said modified restoration signature was received, said restoration signal including the identity of said CHOOSER node, an INDEX value equal to the INDEX value of said modified restora-tion signal, and a predetermined initial REPEAT
value;
e) upon receipt of a complement restoration signal at a TANDEM node:
i. operatively connecting the port upon which said complement restoration signal was received with the port on which its cor-responding restoration signal was received whereby to permit communications traffic therebetween; and ii. incrementing the repeat value of said complement restoration signal and transmit-ting the modified complement restoration signal back through the spare circuit along which said corresponding restoration signal was received;
and f) upon receipt of a modified complement restoration signal at said SENDER node, operatively connecting the port containing communications traffic intended for one of said failed circuits to the port on which said modified complement signature was received.

21. A method as defined in Claim 20, including the step of identifying spare circuits at said one node and allocating a distinct restoration signal to each identified spare circuit.

22. A method as defined in claim 21, wherein said step of designating one of said pair of nodes a SENDER node and the other of said pair of nodes a CHOOSER node being on the basis of an ordinal test conducted independently at each said SENDER and CHOOSER nodes.

23. A method as defined in claim 22, further including the step of storing, at each said SENDER, TANDEM and CHOOSER nodes, received and transmitted signals in respective received and transmitted signal storage registers and the identity of the port of upon which each said signals was received and transmitted.

24. A method as defined in claim 23, wherein step c) further includes the step of ignoring any restoration signal whose REPEAT value is greater than a predeter-mined value.

25. A method as defined in claim 24, wherein step d) further includes the step of ignoring any modified restoration signal whose REPEAT value is greater than a predetermined value.

26. A method as defined in claim 25, wherein step c)(ii) further includes the steps of ignoring a restoration signal having the same INDEX value and the same or greater REPEAT value as that of any restoration signal said node is already transmitting and substitut-ing a restoration signal having the same INDEX value and a smaller REPEAT value as that of any restoration signal said node is already transmitting.

27. A method as defined in claim 26, step e) further including the step of suspending transmission of all modified restoration signals having the same INDEX value as that of said received complement restoration signal.

28. A method as defined in claim 27, each said signal further including a Node Identifier field for identify-ing the node originating a signal.

29. A method as defined in claim 28, further including the step of encoding said restoration and complement restoration signals into a communications traffic signal.

30. An apparatus for restoring communications between a pair of nodes in a network having an arbitrary number of nodes and an arbitrary number of spans interconnect-ing said nodes, each said span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits between nodes capable of but not designated, for transmitting actual communications traffic, said apparatus comprising:
a) means for establishing one or more independent communication paths between said pair of nodes through a series of spare circuits of spans interconnecting said pair of nodes and other interconnected nodes in said network; and b) means for redirecting communications traffic intended for one or more failed spans intercon-necting said pair of nodes through one or more of said paths.

31. A method as defined in Claim 30, said establishing means including:
a) means for repeatedly transmitting restoration signals along logical spans of spare circuits departing said one node;
b) means for repeatedly retransmitting said restora-tion signals from said other interconnected nodes along logical spans of spare circuits departing said interconnected nodes; and c) means at said other node responsive to receipt of a restoration signal thereat for transmitting a complement restoration signal from said other node along a path consisting of the same spare circuits along which said restoration signal was communi-cated to said other node.

32. A method as defined in Claim 31, further including means for identifying spare circuits at said one node and allocating a distinct restoration signal to each identified spare circuit.

33. An apparatus as defined in Claim 32, said trans-mitting means including means for transparently modulat-ing said signals into communications traffic signals travelling between nodes of said network.

34. An apparatus as defined in Claim 31, said repea-tedly transmitting means being operable for repeatedly transmitting said signals along each logical span of spare circuits departing said one node up to the minimum of either the number of circuit restorations required or the number of spare circuits available to said one node.

35. An apparatus as defined in Claim 31, said comple-ment signal transmitting means being operable to respond to restoration signals until the number complement restoration signals equals the number of failed working circuits or until no more spare circuits are available at said other node of said pair of nodes.

36. An apparatus as defined in Claim 31, each said restoration signal and its complement restoration signal including an attribute representative of a particular one of said paths.

37. An apparatus as defined in Claim 31, each said signal including an attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path.

38. An apparatus as defined in Claim 31, each said restoration signal and its complement restoration signal including a first attribute representative of a par-ticular one of said paths and a second attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path.

39. An apparatus as defined in Claim 37, each said restoration signal and its complement restoration signal including third and fourth attributes identifying said one and said other of said pair of nodes.

40. An apparatus as defined in Claim 33, each said restoration signal and its complement restoration signal including a fifth attribute representative of a network-wide identification of the node originating said signals.

41. An apparatus as defined in Claim 31, further including:
means for designating one of said nodes as a SENDER
node from which said restoration signals will originate and the other of said pair of nodes as a CHOOSER from which said complement restoration signals will origi-nate.

42. An apparatus as defined in Claim 40, wherein said designating means being operable to designate said nodes on the basis of an ordinal test conducted independently at each said SENDER and CHOOSER nodes.

43. An apparatus for restoring communications between a pair of nodes in a network having an arbitrary number of nodes and an arbitrary number of spans interconnect-ing said nodes, each said span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits between nodes capable of, but not designated, transmitting actual communications traffic, said method comprising the steps of:
a) means for designating one of said nodes of said pair of nodes as a SENDER node and the other of said pair of nodes as a CHOOSER node;
b) means for establishing one or more independent communication paths between said SENDER node and said CHOOSER node through a series of spare circuits of spans interconnecting said SENDER and CHOOSER nodes with other nodes in said network, said means including:
i. means for repeatedly transmitting restoration signals along each logical span of spare circuits departing said SENDER node up to the minimum of either the number of circuit restorations required or the number of spare circuits available to said SENDER node;
ii. means for repeatedly retransmitting said restoration signals from said other nodes along logical spans of spare circuits departing said other nodes; and iii. means responsive to receipt of a restoration signal at said CHOOSER node for transmitting a complement restoration signal from said CHOOSER node along the logical span of the spare circuit upon which said restoration signal was received;
iv. means for retransmitting said complement restoration signals from said other nodes along the logical span of the spare circuit on which said restoration was received and for connecting the port upon which said restoration signal was received with the port upon which said complement restoration signal was received whereby to permit communications traffic therebetween;
each said restoration signal and its complement restoration signal including a first attribute representative of a network-wide identifier for the node originating said signals, a second attribute identifying either said SENDER node or said CHOOSER node as a SOURCE node, a third attribute identifying either said CHOOSER node or said SENDER node as a TARGET node, a fourth attribute representative of a particular one of said paths, and a fifth, incrementable attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path; and c) means for redirecting, through one or more of said paths, communications traffic received at input ports of each of said SENDER and CHOOSER nodes and intended for transmission through one or more failed spans interconnecting said SENDER and CHOOSER nodes.

44. A method as defined in Claim 43, further including means for identifying spare circuits at said SENDER node and allocating a distinct restoration signal to each identified spare circuit.

45. An apparatus as defined in claim 44, wherein the restoration signals transmitted along said spare circuits departing said SENDER node differ only by the value of said fourth attribute.

46. An apparatus as defined in claim 44, wherein said retransmitting means being operable to increment said fifth attribute of received signals prior to retrans-mitting said signals.

47. An apparatus as defined in claim 44, further including means at said other nodes and said CHOOSER
node for rejecting any received signal whose fifth attribute exceeds a predetermined value.

48. An apparatus as defined in claim 44, storage register means at each other node for storing the first restoration signal received in connection with a particular path and rejecting any other received restoration signals relating to said particular path.

49. An apparatus as defined in claim 44, storage register means for storing at each other node the first restoration signal received in connection with a particular path, and means for rejecting any other received restoration having the same first to fourth attributes but a larger fifth attribute and for replac-ing said stored signal with any new restoration signal having the same first to fourth attributes but a smaller fifth attribute.

50. An apparatus for restoring communications between a pair of nodes in a network having an arbitrary number of nodes and an arbitrary number of spans interconnect-ing said nodes, each said span having working circuits between nodes designated for transmitting actual communications traffic and spare circuits between nodes capable of, but not designated, transmitting actual communications traffic, said method comprising the steps of:
a) means responsive to a communication failure in one or more circuits in one or more spans between said pair of nodes for designating one of said pair of nodes a SENDER node and the other of said pair of nodes a CHOOSER node;
b) means for identifying spare circuits at said SENDER node and allocating a distinct restoration signal to each identified spare circuit;
c) means for repeatedly transmitting restoration signals on the outgoing transmission link of each said identified spare circuits departing from said SENDER node, each said restoration signal includ-ing the identity of said SENDER node, the identity of said CHOOSER node, an INDEX value representa-tive of the identity of a complete, independent route between said SENDER and said CHOOSER nodes starting at the spare circuit departing said SENDER node along which said signal is trans-mitted, and a predetermined initial REPEAT value;
d) means responsive to receipt of one of said restoration signals at a TANDEM node for determin-ing whether said TANDEM node is already transmitt-ing a restoration signal having the same INDEX
value as the just received restoration signal, ignoring any restoration signal having the same INDEX value as that of a signal said node is already transmitting and, if said node is not already transmitting a restoration signal having the same INDEX value, producing a modified restoration signal by incrementing said repeat value of said restoration signal and repeatedly transmitting said modified restoration signal along spare links departing said TANDEM node;
e) means responsive to a modified restoration signal at said CHOOSER node for repeatedly transmitting a complement restoration signal from said CHOOSER
node back through the same spare circuit along which said modified restoration signature was received, said restoration signal including the identity of said CHOOSER node, an INDEX value equal to the INDEX value of said modified restora-tion signal, and a predetermined initial REPEAT
value;
f) means responsive to a complement restoration signal at a TANDEM node for operatively connecting the port upon which said complement restoration signal was received with the port on which its corresponding restoration signal was received whereby to permit communications traffic therebet-ween, incrementing the repeat value of said complement restoration signal and transmitting the modified complement restoration signal back through the spare circuit along which said corresponding restoration signal was received; and g) means responsive to a modified complement restora-tion signal at said SENDER node, operatively connecting the port containing communications traffic intended for one of said failed circuits to the port on which said modified complement signature was received.

51. An apparatus as defined in claim 50, said desig-nating means for designation one of said pair of nodes a SENDER node and the other of said pair of nodes a CHOOSER node being operable on the basis of an ordinal test.

52. An apparatus as defined in claim 51, further including storage register means at each said SENDER, TANDEM and CHOOSER nodes, for storing received and transmitted signals and the identity of the port of upon which each said signals was received and transmitted.

53. An apparatus as defined in claim 52, wherein said restoration signal responsive means being operable to ignore any restoration signal whose REPEAT value is greater than a predetermined value.

54. An apparatus as defined in claim 53, wherein said modified restoration signal responsive means being operable to ignore any modified restoration signal whose REPEAT value is greater than a predetermined value.

55. An apparatus as defined in claim 54, wherein said restoration signal responsive means being operable to ignore a restoration signal having the same INDEX value and the same or greater REPEAT value as that of any restoration signal said node is already transmitting;
and substitute a restoration signal having the same INDEX value and a smaller REPEAT value as that of any restoration signal said node is already transmitting.

56. An apparatus as defined in claim 55, said modified complement restoration signal responsive means being further operable to suspend transmission of all modified restoration signals having the same INDEX value as that of said received complement restoration signal.

57. An apparatus as defined in claim 56, each said signal further including a Node Identifier field for identifying the node originating a signal.

58. An apparatus as defined in claim 57, further including the step of encoding said restoration and complement restoration signals into a communications traffic signal.

59. In a communications network having an arbitrary number of nodes and an arbitrary number of spans interconnecting said nodes, each said span having working circuits between nodes designated for transmitt-ing actual communications traffic and spare circuits between nodes capable of but not designated for the transmission of actual communications traffic, each said node having one or more bi-directional transmission interfaces connected to external transmission lines and to an internal switching matrix, each said transmission interface having circuit means for processing signals received along a receive link and feeding said received signal to said switching matrix and circuit means for processing transmit signals received from said switching matrix and applying said transmit signal to a transmit link, the improvement comprising a communications restoring apparatus for use at each said node and comprising:
restoration signature detecting circuit means at each said transmission interface for detecting and storing restoration signatures received along said receive link;
restoration signature transmitting circuit means at each said transmission interface for applying transmit signature signals to said transmit link; and control means operatively connected to said detecting circuit means and said transmitting circuit of each said transmission interface and being responsive to:
a) an alarm signal indicative of a circuit failure in a span connecting said node and an adjacent node for generating and repeatedly applying predeter-mined restoration signature signals to said restoration signature transmitting circuit means of one or more of said transmitting interfaces whereby to cause said restoration signals to be transmitted along one or more transmitting links to other adjacent nodes;

b) a restoration signature detected by said detecting circuit for producing modified restoration signatures and applying said modified restoration signatures to the transmitting circuit means of each said transmission interface of said node whereby to cause said modified restoration signatures to be transmitted along said transmitt-ing links to adjacent nodes;
c) a modified restoration signature received at the detecting circuit means of one of said transmis-sion interfaces for generating a complement restoration signature signal and applying said complement restoration signature signal to said transmitting circuit means of said one of said transmission interfaces; and d) a complement restoration signature received at the detecting circuit means of one of said transmis-sion interfaces for either operatively connecting the receive and transmission links of the trans-mission interface at which said complement restoration signature signal was received with the transmission and receive links, respectively, of the transmission interface on which the restora-tion signature associated to said complement signal was received if said control means modified an existing restoration signature, or operatively connecting the receive and transmission links of the transmission interface at which said comple-ment restoration signature signal was received with the transmission and receive links, respec-tively, of the transmission interface on which communications traffic intended to be transmitted through a failed circuit between said pair of nodes if said control means generated the original restoration signature associated with said complement restoration signature signal.

60. An apparatus as defined in claim 59, said signa-ture receiving circuit means including:
a signature detector circuit for recovering restoration signature attributes transparently encoded in a signal received on said received link and producing an output representative of said attributes;
a signature reception register adapted to receive said detector circuit output;
a receive signature storage register for storing attributes of a restoration signature; and a validity and logic control circuit means for comparing the contents of said signature reception register with the contents of said signature storage register and transferring the contents of the former into the latter and transmitting a flag to said control means that a new signature has been received.

61. An apparatus as defined in claim 59, said signa-ture transmitting circuit means including:
a transmit signature storage register for receiving and storing a transmit restoration signature;
a transmit signature modulation circuit for modulating the contents of said transmit signature storage register into a transmit signal applied to said transmit signal processing circuit means; and control logic means responsive to said control means for loading a new signature into said transmit signature Storage register and, once loaded, causing repeated circulation of the contents of the register whereby to cause said transmit restoration signature to be indefinitely, repeatedly applied to said modulation circuit.

62. An apparatus as defined in claim 59, further including:
Interface port number register means for storing the number of the port associated with said signature receiving and transmitting circuit means; and a port status register for storing predetermined port status information.

63. An apparatus as defined in claim 62, further including circuit failure alarm detection means as-sociated with each said transmission interface and responsive to failures in said transmission and receive links of said associated transmission interface whereby to cause the contents of said port number register and said port status register to be transmitted to said control means.

64. An apparatus as defined in claim 60, said signa-ture transmitting circuit means including:
a transmit signature storage register for receiving and storing a transmit restoration signature:
a transmit signature modulation circuit for modulating the contents of said transmit signature storage register into a transmit signal applied to said transmit signal processing circuit means; and control logic means responsive to said control means for loading a new signature into said transmit signature Storage Register and, once loaded, causing repeated circulation of the contents of the register whereby to cause said transmit restoration signature to be indefinitely, repeatedly applied to said modulation circuit.

65. An apparatus as defined in claim 64, further including:
interface port number register means for storing the number of the port associated with said signature receiving and transmitting circuit means; and a port status register for storing predetermined port status information.

66. An apparatus as defined in claim 65, further including:
circuit failure alarm detection means associated with each said transmission interface and responsive to said transmission and receive links of said as-sociated transmission interface whereby to cause the contents of said port number register and said port status register to be transmitted to said control means.

67. An apparatus as defined in claim 64, each said restoration signal and its complement restoration signal including a first attribute representative of a network-wide identifier for the node originating said signals, a second attribute identifying either a SENDER node or a CHOOSER node as a SOURCE node, a third attribute identifying either said CHOOSER node or said SENDER node as a TARGET node, a fourth attribute representative of a particular one of said paths, and a fifth, incrementable attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path, each said receive and transmit storage registers being adapted to store each said attributes of each received restoration signal and transmit complement restoration signal, respectively.

68. An apparatus as defined in claim 65, said Port Status Register being adapted to store a first signal indicative of whether receive and transmit links associated with said port status register is currently experiencing loss of transmission integrity, a second signal indicative of whether the circuit associated with said port is a working circuit or a spare circuit, a third signal indicative of whether said port is con-nected to another port of said node and a fourth signal representative of the identity of another port to which said port may be connected.

69. A Digital Crossconnect Machine for use as a node in a communications network having an arbitrary number of nodes and an arbitrary number of spans interconnect ing said nodes, each said span having one or more working circuits interconnecting adjacent nodes and designated for transmitting actual communications traffic and spare circuits interconnecting adjacent nodes and capable of, but not designated, for the transmission of actual communications traffic, each said circuit having a transmit link and a receive link, said machine comprising:
an operating system for controlling the operation of said machine;
crossconnect switching matrix means for internally connecting interface ports of said machine;
crosspoint controller means responsive to said operating system for controlling said switching matrix means;
one or more bi-directional transmission interfaces connected to external transmission lines and to an internal switching matrix each said transmission interface having:
circuit means for processing signals resolved along the receive link of an associated working or spare circuit and feeding said received signal to said switching matrix;
circuit means for processing transmit signals received from said switching matrix and applying said transmit signal to a transmit link;
restoration signature detecting circuit means for detecting and storing restoration signatures applied to said received signal processing circuit means along said receive link; and restoration signature transmitting circuit means for applying transmit signature signals to said transmit signals processing circuit means for transmission along transmit link; and an emergency restoration controller operatively con-nected to each said transmission interface means, said crosspoint controller and said operating system and being responsive to a circuit failure alarm and/or restoration signature detected by said detecting circuit and cooperable with other Digital Crossconnect Machines in said network for establish-ing one or more independent communication paths between said pair of nodes through a series of spare circuits of spans interconnecting a pair of nodes between which communications have failed and other interconnected nodes and redirecting communications traffic intended for one or more failed spans through said one or more of said paths.

70. A Digital Crossconnect machine as defined in claim 69, each said signature receiving circuit means including:
a signature detector circuit for recovering restoration signature attributes transparently encoded in a signal received on said received link and producing an output representative of said attributes;
a signature reception register adapted to receive said detector circuit output;
a receive signature storage register for storing attributes of a restoration signature; and a validity and logic control circuit means for comparing the contents of said signature reception register with the contents of said signature storage register and transferring the contents of the former into the latter and transmitting a flag to said control means that a new signature has been received.

71. An Digital Crossconnect Machine as defined in claim 69, each said signature transmitting circuit means including:
a transmit signature storage register for receiving and storing a transmit restoration signature;
a transmit signature modulation circuit for modulating the contents of said transmit signature storage register into a transmit signal applied to said transmit signal processing circuit means; and control logic means responsive to said control means for loading a new signature into said transmit signature Storage Register and, once loaded, causing repeated circulation of the contents of the register whereby to cause said transmit restoration signature to be indefinitely, repeatedly applied to said modulation circuit.

72. A Digital Crossconnect Machine as defined in claim 69, each said transmission interface means further including:
an interface port number register for storing the number of the port associated with said signature receiving and transmitting circuit means; and a port status register for storing predetermined port status information.

73. A Digital Crossconnect Machine as defined in claim 69, each said transmission interface means further including:
circuit failure alarm detection means associated with each said transmission interface and responsive to said transmit and receiving links of said associated transmission interface whereby to cause the contents of said port number register and said port status register to be transmitted to said control means.

74. A Digital Crossconnect Machine as defined in claim 70, each said signature transmitting circuit means including:
a transmit signature storage register for receiving and storing a transmit restoration signature;
a transmit signature modulation circuit for modulating the contents of said transmit signature storage register into a transmit signal applied to said transmit signal processing circuit means; and control logic means responsive to said control means for loading a new signature into said transmit signature Storage Register and, once loaded, causing repeated circulation of the contents of the register whereby to cause said transmit restoration signature to be indefinitely, repeatedly applied to said modulation circuit.

75. A Digital Crossconnect Machine as defined in claim 74, each said transmission interface means further including:
an interface port number register for storing the number of the port associated with said signature receiving and transmitting circuit means; and a port status register for storing predetermined port status information.

76. A Digital Crossconnect Machine as defined in claim 76, each said transmission interface means further including:
circuit failure alarm detection means responsive to said transmission and receive links of said associated transmission interface whereby to cause the contents of said port number register and said port status register to be transmitted to said control means.

77. An apparatus as defined in claim 76, each said restoration signal and its complement restoration signal including a first attribute representative of a network-wide identifier for the node originating said signals, a second attribute identifying either a SENDER node or a CHOOSER node as a SOURCE node, a third attribute identifying either said CHOOSER node or said SENDER node as a TARGET node, a fourth attribute representative of a particular one of said paths, and a fifth, incrementable attribute representative of the number of spans and/or linking nodes in each said path at any given point along said path, each said receive and transmit storage registers being adapted to store each said attributes of each received restoration signal and transmit complement restoration signal, respectively.

78. An apparatus as defined in claim 77, said Port Status Register being adapted to store a first signal indicative of whether receive and transmit links associated with said port status register is currently experiencing loss of transmission integrity, a second signal indicative of whether the circuit associated with said port is a working circuit or a spare circuit, a third signal indicative of whether said port is con-nected to another port of said node and a fourth signal representative of the identity of another port to which said port may be connected.