CA1261442A - Nonblocking switching system and method - Google Patents
Nonblocking switching system and methodInfo
- Publication number
- CA1261442A CA1261442A CA000522591A CA522591A CA1261442A CA 1261442 A CA1261442 A CA 1261442A CA 000522591 A CA000522591 A CA 000522591A CA 522591 A CA522591 A CA 522591A CA 1261442 A CA1261442 A CA 1261442A
- Authority
- CA
- Canada
- Prior art keywords
- network
- sources
- destinations
- paths
- active
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q3/00—Selecting arrangements
- H04Q3/64—Distributing or queueing
- H04Q3/68—Grouping or interlacing selector groups or stages
Abstract
NONBLOCKING SWITCHING SYSTEM AND METHOD
Abstract of the Disclosure A NONBLOCKIG switching system for communications systems is characterized by two identical blocking networks, each for being connected in series between sources and destinations to be selectively interconnected. One network actively established paths between the sources and destination while the other, although conditioned to be able to provide such paths, remains inactive in a standby mode for establishing the paths should the active network fail. If a blocking state occurs in the active network before its maximum realizable capacity is reached, a nonblocking condition of the system is established by eliminating all of the paths in the inactive network, and then reestablishing the paths in accordance with a relatively simple algorithm that loads the network without a blocking state being reached before the network is completely full. The roles of the active and inactive networks are then reversed, so that the previously inactive network becomes active and establishes the paths between sources and destinations and the previously active network becomes inactive. the paths in the newly active network are then copied into the newly inactive network so that it will be prepared to connect the sources and destinations should the active network fail.
Abstract of the Disclosure A NONBLOCKIG switching system for communications systems is characterized by two identical blocking networks, each for being connected in series between sources and destinations to be selectively interconnected. One network actively established paths between the sources and destination while the other, although conditioned to be able to provide such paths, remains inactive in a standby mode for establishing the paths should the active network fail. If a blocking state occurs in the active network before its maximum realizable capacity is reached, a nonblocking condition of the system is established by eliminating all of the paths in the inactive network, and then reestablishing the paths in accordance with a relatively simple algorithm that loads the network without a blocking state being reached before the network is completely full. The roles of the active and inactive networks are then reversed, so that the previously inactive network becomes active and establishes the paths between sources and destinations and the previously active network becomes inactive. the paths in the newly active network are then copied into the newly inactive network so that it will be prepared to connect the sources and destinations should the active network fail.
Description
Back~und of tlle Invention The present invention relates to switching systems for communications systems, and in psrticular to an improved switching system which includes a pair of three-stage connecting networks and a method of operating the same. ' Many communicntions systems, particularly telephone -systems, contain a connecting network. A connecting -, network is an arrangement of switches and transmission links through which terminals connected to sources and terminals connected to destinations may be connected .-together in many combinations. --Calls in progress in a connecting network do not arise in a selected order or time sequence. Requests '^ :~
for connections (new calls) and terminations of connections (hangups) occur more or less at random, and --the performance of a network when subjected to random traffic is measured by the fraction of requested connections that cannot be completed, or by the probability of blocking.
The performance of a connecting network for a given level of traffic is determined largely by its structure, i~
i.e., by what terminals have a switch placed between them and can be connected together by closing the switch. The structure of a network determines what ;, combinations of terminals can be connected together 5 simultaneously, and if the structure is too simple, only B relatively few calls can be in progress at the same time. lf the structure is extensive and complex, it may accommodate a large number of simultaneous calls in i--progress, but the network may be expensive to buiId and difficult to control.
~ ' 4,~
Cnnn~cting networks made of square switches have long been used in telephone central offices. In a network made of square switches, each switch hss the same number of inlets as outlets. The network is said to be arranged in stages if the switches are partitioned into sets called stages, two of which respectively carry the inlets to and outlets from the network, while the others are internal stages between the two, with the arrangement being such thHt every psth from an inlet to an outlet passes through each stage only once. If the ;
network is full access, each switch in a given stage has a link to every switch in the adjacent StQgeS. '~
A network is in a blocking state if there are an idle inlet to and outlet from the network, but no-available path exists through the network to connect the -inlet and outlet. A routing algorithm may be used for -choosing routes through the network for unblocked calls during its operation, or in the alternative to -reestablish paths through the network in a manner to --prevent the network from going to a blocking state. g,r`' Such a network is said to be nonblocking in the wide .-sense if it is capable of blocking, but there is an a]gorithm for assigning routes to incoming CRI15 that --precludes any blocking, such that if the rule is ~ -applied, no blocking state is accessible. The system csn then satisfy demands for connections as they arise without rejecting any or rerouting existing calls.
However, etworks controlled by such algorithms and the algorithms themselves are often complex.
A connecting network may a]so be nonblocking in the strict sense, if no matter what state the network may be ~~~
in, it is always possible to connect together an idle ``'' 4'-`
pair of terminnls without disturbing calls nlresdy in progress. In most new telephone switching systems, nonblocking in the strict sense is schieved by an N2 type of network. An example of this would be a 2000 port network which requires 4 million switch elements to ensure nonblocking, but such networks are quite complex and extensive and add considersble cost to the system.
Objects of the Invention An object of the present invention is to provide an improved switching system which is reduced in complexity and cost, but in which blocking states may easily be QVO i ded. ~-Another object is to provide such a switching -system, which comprises two connecting networks for reliability through redundancy, but by virtue of which redundancy the networks may conveniently be made :'t ~ ' nonblocking should a blocking state occur.
A further object is to provide a method of operating such a switching system.
Summary of the Invention In accordance with the present invention,~there is provided a switching system for connecting individual --~
ones of a plurality of sources to individual ones of a -~
plurslity of destinations in various combinations. The ~,~
f system comprises a peir of connecting networks that are `-`
! ~
nonblocking in the wide sense, each for being connected ;}, in series between all of the sources and Bll of the ~ -destinations. Also included are means for controlling ~.
the networks to estsblish and interrupt paths therethrough between individual ones of the sources and ~.i destinations, along with selector means for connecting ~ -one or the other of the networks in series between the -,ji~:
44~
sources ~nd destinntions, such thnt the network connected in series is active to establish connections between the sources and destinations, whi]e the other network is inactive. The arrsngement is such that the means for controlling, upon occurrence of A blocking state in the active network before its maximum capacity is reached and the network is full, includes means for `.
emptying the inactive network of all paths therethrough and then establishing paths through the network which will connect the sources and destinations in the manner attempted when the active network went to a blocking state -- without occurrence of a blocking state in the inactive network. After the paths are established through the inactive network, the selector means is operative to connect the inactive network in series between the sources and destinations and to disconnect the active network from in series, whereby the roles of -the networks are reversed and the previously inactive network becomes active while the previously active -network becomes inactive.
The invention also contemplates a method of n connecting individual ones of a plurality of sources to ~~
individual ones of a plurality of destinations in -~
various combinations. The method comprises the steps of . coupling all of the sources to inlets to each of first -`
and second connecting networks that are nonblocking in .-the wide sense, coupling all of the destinations to 'J'~
outlets from each of the networks, and connecting the first network only in series between the sources and `--destinations. Paths are then established and interrupted through the first network to connect and disconnect individual ones of the sources and ~
destinations. ~pon occurrence of 8 blocking stnte in the first networ~ before it has reached its maximum capacity and is full, paths are established through the second network which will connect the sources and -~
destinations in the manner attempted when the blocking state occurred in the first network -- without causing a blocking state to occur in the second network. The seeond network is then conneeted in series between the sources and destinations and the first network is diseonneeted from in series, whereby the seeond network ~
then eonneets the sources and destinations. ~-The foregoing and other objects, advantages and features of the invention will become apparent upon a consideration of the following detailed deseription, when taken in eonjunetion with the aeeompanying ;~
drawings.
Brief Deseription of the Drawings Fig. 1 is a bloek diagram of a known type of digital switching system, illustrating basie components of the system, whieh inelude a digital eonneeting `~
network;
Fig. 2 is a simplified bloek diagram of a digital , switching system according to the teachings of the present invention;
Fig. 3 is a block diagram of a 4 x 4 connecting network of a type which may represent each network of Fig. 2; and ~ ig. 4 is similar to Fig. 3, except that it illustrates A n x n connecting network. ~
~.~'. .
;..
~ ' ~.144~
Detailed Description The Prior Art Fig. 1 shows in block diagram form a typical'prior art digital switching system ineluding a line and trunk -interface, a digital eonnecting network and a eentra]
control. The line and trunk interface eonneets the system to the outside world, e.g., to serviee eireuits, --i digital trunks, analog trunks and subseriber lines.
.. ...
Where the signals from the outside world are analog, the ~
" .
interfaee eonverts them to digital.
The digital eonneeting network is an arrangement of switehes snd transmission links through whieh terminals connected to sourees of signals may be seleetively coupled in many eombinations to terminals eonneeted to '-destinations, the basie funetion of the digital conneeting network being to perform the neeessary ~--switehing to estsblish the paths for the conneetions.
The eentral eontrol in turn provides stored program ~^~
eontrol for both the syseem and a man/maehine interfaee (not shown). ~--Calls in progress in a digital eonneeting network ~ -~
do not usually arise in a predetermined order or time sequenee. Requests for eonnections (new ealls) and terminations of connections (hangups) occur more or less -~
at random, and unless the network is nonbloeking in the strict sense, or nonbloeking in the wide sense, it can ~-~
happen that the it is not able to establish a connection -~
between a calling source and a selected destination.
~hen that occurs, the network is said to be in a blocking state, i.e., there is an idle inlet to and an id]e out]et irom the network, but there is no availab]e psth ~hrough the network to connect the inlet to the ;:
out]et 7 To AVO id a blocking condition, most new telephone switching s~stems use connecting networks that are nonblocking in the strict sense. However, a strictly nonblocking network is often extensive and complex since nonblocking is normally schieved by an N2 type of network, such that a 2000 port network requires 4 million switch elements to ensure nonblocking.
The Invention -~
~o mitigate cost and eomplexity- disadvantages of conventional switching systems having eonneeting -~
networks that are nonbloeking in the striet sense or :.
nonbloeking in the wide sense, RS seen in Fig. 2 the -~
invention provides an improved switehing system which ineludes a pair of identical conneeting networks 0 and 1, each of whieh is nonblocking in the wide sense.
Every source is connected through each network to every destination, snd elthough only 8 single source and i---destination are illustrated for simplicity, it is -`~
understood that in aetual use, the networks would be conneeted between pluralities of sourees and destinations. -~ -~ . ~
At any given time, only one network aetively establishes paths from sourees to destinations as -~
determined by and through a conventional 1 of 2 selector ;-circuit. The other network is inactive at the time and in a standby mode, although the same paths established -~
through the active network Q150 exist in the inactive ,;
network. This type of redundant configuration of networks increases the reliability of the switching ,~
svstem, such that if the active network fails, the other ~-one will be prepared to maintain the paths and assume switching functions. However, according to the .:.
~t.~44~`
invention, the redundnnt contigurstion also is used to estab]ish an easily implementable nonblocking environment.
The connecting networks O and 1 are n x n three stage networks which are responsive to a central control having a man/machine interfMce (not shown). Connections are added and dropped in a random manner, so blocking can occur. When a blocking state occurs in the active .
network before it has reached its maximum capacity and is full, instead of using a complicated algorithm to .
rearrange existing connections in the active network, -.
according to the ir,vention a nonblocking condition of -the system is easily reached by: (1) eliminating all of the paths in the inactive network, so that the network .
is empty; (2) reestablishing paths in the inactive -~
network, according to a relatively simple algorithm, that will connect the sources to the destinations in the manner attempted when the active network reached R
blocking state, without a blocking state occurring in ~-the inactive network until the network is full; (3) ~-.
switching or reversing the roles of the active and inactive networks, so that the previously inactive _ -network becomes active and establishes the paths from - -the sources to the destinations through the 1 of 2 selector circuit, and the previously active network becomes inactive; and then (4) copying the paths established in the newly active network into the newly ~-~
inactive network, so that the inactive network will be prepared to connect the sources to the destinations and `~
assume switching functions should the active network ~=.-fail. ~
~'~
~ 4, To demonstrate thRt paths may be established throu@h an initially empty n x n three stage network, such thst no b]ocking state occurs unti] the maximum realizable capacity of the network is reached and the network is full, consider the 4 x 4 three stage network shown in Fig. 3, which may represent each of the networks 0 and 1. The network is comprised of square switches, eflch having the same number of inlets AS -~
outlets. ~or a 4 x 4 network, each stage has 4 switches and each switch has 4 inlets and 4 outlets, so inputs Yo-Y3 each represent four lines from respective sources to associated input terminals of source switches So-S3, and outputs Xo-X3 each represent four lines to respective destinations from associated output terminals --of destination switches Do-D3. The network is full àccess, so each switch in a given stage has a transmission link to every switch in adjacent stages, -although for simplicity of illustration only links between the source switch So and intermediate stage ,-switches 10-13, as well as between the destination switch Do and the intermediate stage switches, have been shown.
Starting with an initially empty network, i.e., whichever network 0 or 1 was inactive and emptied of -paths when the other became blocking, to connect the sources and destinations through the inactive network in the manner attempted when the active network reached a blocking state, such that the inactive network does not ~~
become b]ocking until it is completely full, a ~-connection from S~ to Dj is established by choosing an ~-intermediate stage Ij, through which the connection is ~-`
to pass, according to the following algorithm: , -In- ,", ~ 14~
.
Choose the path (Dj, lo), (Dj, Il), (Dj, 12) or (Dj, 13) that has the minimum number of existing connections on that particular link.
After the path is chosen and established, the psth (1;, Sk) is established to complete the connection from Sk to Dj.
The above path choosing and establishing seguence is then repeRted until all of the desired connections that are to exist between sources and destinations have been established. The 1 of 2 selector circuit is then operated to reverse the roles of the active and inactive -networks, such that the previous]y inactive network, with the newly established paths therein, becomes active -and connects the sources to the destinations, and the previously sctive network becomes inactive. The paths in the newly active network are then copied into the newly inactive network, so that it will be prepared to assume connection and switching functions should the sctive network fail.
To appreciate that connections may be Qdded to the ~-emptied insctive network, so thst no blocking occurs until the network is full, considering the following:
Note 1: The capacity of link (Dj, Ij) is not exceeded -~
since each link hHs a maximum capacity of Xmax/4 and Xj/4 < XmaX/4~ where Xj is the traffic leaving Dj and XmaX is the maximum traffic capacity of Dj.
Note 2 Since Xo < Xmax~ , X3 < XmaX~ the maximum traffic through the center stage li is ,~
Xo/9 + X1/4 + X2/4 + X3/4 < 1/4(4XmaX) = Xmax .i-Therefore, the maximum capacity of the center ~-stage is not exceeded.
.
- 1 ~ - ....
~t;144~
~ote 3: Lel tj k = (the number of connections from D
to S~ max~ or the portion of Xj that is connected to Sk, and < tj k < 1. Then, the -~
maximum traffic over link (li, Sk) =
1/4XûtOk + 1/4Xltlk + 1/4X2t2k + 1/4X3t3k , < 1/4Xmax (tOk + t1k + t2k + t3k). ~owever, for a ful] network Xmax(tok + t1k + t2k + t3k) ; = XmaX~ since the maximum traffic entering Sk .i from the left side of the network is also YmaX :i;
= XmaX~ Therefore, for a full network (tOk +
....
tlk ~ t2k + t3k) = 1, although in general, and to include the situation where the network is less than full, (tOk + t1k + t2k + t3k) < 1- . -Thus, the maximum traffic on link (Ij, Sk) -~ -< 1/4 (Xmax). This satisfies the maximum link -requirements of link (1;, Sk).
As seen from note 1, the (Dj, Ij) inter~ediate link capacity is not exceeded; according to note 2, the ~:
center stage capacity is not exceeded; and from note 3 it is seen that the (Ij, Sk) link capacity is not -exceeded. The network therefore remains nonb]ocking when connections are sequentially added to the initially ~-empty network according to the above algorithm.
The proof that paths may be established through an ' `
initially empty three stage network, such that the network remains nonblocking until it is full, may be '~
extended to any n x n network. ~ -Consider the full access n x n network in Fig. 4, `
which may represent each of the networks O and 1.
Starting with the network initiall~ empty, to establish ;~`
B connection from Sk to Dj, choose an intermediate stage ,~
-12- ~
144~, Ij, tllrollgh which the connection is to pRSS, ~ccording to th~ fo~lowing algoritl)m:
Choose the path [(Dj, lo), (D~ ), . . ., or (Dj, In_l)] that has the minimum number of existing connections on that particular link.
After the path is chosen and established, the path (Ij, Sk) is established to complete the connection from Sk to Dj.
The above path choosing and establishing sequence ' -is then repeated until all oi the desired connections that are to exist between sources and destinations have ..
been established. The 1 of 2 selector circuit is then operated to reverse the roles of the active and inactive .:
networks, such that the previously-inactive network, ,'-:
with the newly established paths therein, becomes active ~-and connects the sources to the destinations, and the previously active network becomes inactive. The paths ,-.
in the newly active network are then copied into the newly inactive network, so that it will be prepared to assume connection and switching functions should the i.
active network fail. :-To appreciate that connections may be added to the '~~
network, so that no blocking occurs until the network is --full, consider the following: .
Note 1: The CQpQCity of link (Dj, I;) is not exceeded since each link hss a maximum capacity of max/n and Xj/n < XmaX/n~ where Xj is the ~ -traffic leaving Dj and XmaX iS the maximum ' traffic capacity of Dj.
Note 2: Since Xo < XmQX~ , Xn-l < Xmax~ the maximum traffic through the center stage Ij is ,.
-13- ~
i~il4~2 ~;n/n + Xl/n ~ Xn_l/n < l/n(nXmax) =
max Therefore, the maximum capac;ty of the center stsge is not exceeded.
~ote 3: Let tj k = (the number of connections from Dj to stage Sk)/XmaX~ or the portion of Xj that is connected to Sk, and 0 < tj k < 1. Then, the mnximum traffic over link (Ij, Sk) =
1/n Xoto k + . . . + 1/n Xjt; k + . . +
l/n Xn_1tn_1 < 1/n Xmax (t0,k + + tj,k +
. . . + tn_1 k). However, for a full network Xmax ~t0,k + + tj,k + . . . + tn_1 k) =
XmaX~ since the maximum traffic entering Sk from the left side of the network is also YmaX
= XmaX~ Therefore, for a full network (to k +
+ tj,k + tn_1,k) = 1, a]though in ,.
general, and to include the situation where the network is less than full, (t0,k + + tj,k + + tn-l,k) < 1. ',~
Again, the maximum traffic on link (li, Sk) < -1/n(Xmax). This satisfies the maximum link requirements `-~
of link (1;, Sk). .r "
The invention therefore contemplates a switching .
system having two identical, full access, n x n, three ~-~
stage blocking networks. One network at a given time is : ~.
r active and connects the sources to the destinations, 1.~
while the other is inactive and in a standby mode, but ~-has in it the same paths as are in the active network, ,.
so that shou]d the active network fail, the insctive one ~, will be ready to assume connection and switching functions. The redundancy provides incressed reliability in the event of failure of one of the networks.
Li~l44;2 l`he redundancy also enables the switching system to ~e mRinlained in fl nonb]ocking mode. Although the networks may reach a blocking state, should that occur it may readily be resolved. In particular, upon occurrence of a blocking stnte in the active network, the inactive network is simply cleared of the paths therein, and new paths are established that will connect the sources and destinations in the manner attempted when the active network became blocking, without e blocking state occurring in the inactive network until it is fully loaded. The roles of the active and inactive networks are then reversed.
By virtue of the arrangement, it is not necessary ~`
to interrupt and rearrange connections through the --active network should it reach a blocking state, so there is no interruption of communicfltions services. In addition, the particular- algorithm for establishing the transmission paths through the inactive network is - -~
reletively simple compared with prior algorithms used to esteblish and rearrange paths through networks that are nonblocking in the wide sense. In essence, efter -;
emptying the inactive network, paths are reestablished through its center stage, beginning with lo end ;
sequentially cycling from lo to In_1, to make the ,~
connections between the center stage switches and the various output stage switches Do to Dn_1. After each , path between a switch of the center stage and a switch , of the output stage is established, one is then ~-established between the particular center stage switch -`:
and a selected switch of the input stage, such that a connection from Sk to Dj is completed. The ru]e for ;-msking the connections from S~ to Dj ensures that the . ~
- 1 5 - r::
144~
network m~y ~c completely lorlded ~o its msximum capacity before it renches n blocking state.
~ 'hile embodiments of the invention have been described in detail, various modificntions and other embodiments thereof may be devised by one skilled in the art without departing from the spirit and scope of the invention, as defined in the appended claims.
~.
. . .
.':
...
for connections (new calls) and terminations of connections (hangups) occur more or less at random, and --the performance of a network when subjected to random traffic is measured by the fraction of requested connections that cannot be completed, or by the probability of blocking.
The performance of a connecting network for a given level of traffic is determined largely by its structure, i~
i.e., by what terminals have a switch placed between them and can be connected together by closing the switch. The structure of a network determines what ;, combinations of terminals can be connected together 5 simultaneously, and if the structure is too simple, only B relatively few calls can be in progress at the same time. lf the structure is extensive and complex, it may accommodate a large number of simultaneous calls in i--progress, but the network may be expensive to buiId and difficult to control.
~ ' 4,~
Cnnn~cting networks made of square switches have long been used in telephone central offices. In a network made of square switches, each switch hss the same number of inlets as outlets. The network is said to be arranged in stages if the switches are partitioned into sets called stages, two of which respectively carry the inlets to and outlets from the network, while the others are internal stages between the two, with the arrangement being such thHt every psth from an inlet to an outlet passes through each stage only once. If the ;
network is full access, each switch in a given stage has a link to every switch in the adjacent StQgeS. '~
A network is in a blocking state if there are an idle inlet to and outlet from the network, but no-available path exists through the network to connect the -inlet and outlet. A routing algorithm may be used for -choosing routes through the network for unblocked calls during its operation, or in the alternative to -reestablish paths through the network in a manner to --prevent the network from going to a blocking state. g,r`' Such a network is said to be nonblocking in the wide .-sense if it is capable of blocking, but there is an a]gorithm for assigning routes to incoming CRI15 that --precludes any blocking, such that if the rule is ~ -applied, no blocking state is accessible. The system csn then satisfy demands for connections as they arise without rejecting any or rerouting existing calls.
However, etworks controlled by such algorithms and the algorithms themselves are often complex.
A connecting network may a]so be nonblocking in the strict sense, if no matter what state the network may be ~~~
in, it is always possible to connect together an idle ``'' 4'-`
pair of terminnls without disturbing calls nlresdy in progress. In most new telephone switching systems, nonblocking in the strict sense is schieved by an N2 type of network. An example of this would be a 2000 port network which requires 4 million switch elements to ensure nonblocking, but such networks are quite complex and extensive and add considersble cost to the system.
Objects of the Invention An object of the present invention is to provide an improved switching system which is reduced in complexity and cost, but in which blocking states may easily be QVO i ded. ~-Another object is to provide such a switching -system, which comprises two connecting networks for reliability through redundancy, but by virtue of which redundancy the networks may conveniently be made :'t ~ ' nonblocking should a blocking state occur.
A further object is to provide a method of operating such a switching system.
Summary of the Invention In accordance with the present invention,~there is provided a switching system for connecting individual --~
ones of a plurality of sources to individual ones of a -~
plurslity of destinations in various combinations. The ~,~
f system comprises a peir of connecting networks that are `-`
! ~
nonblocking in the wide sense, each for being connected ;}, in series between all of the sources and Bll of the ~ -destinations. Also included are means for controlling ~.
the networks to estsblish and interrupt paths therethrough between individual ones of the sources and ~.i destinations, along with selector means for connecting ~ -one or the other of the networks in series between the -,ji~:
44~
sources ~nd destinntions, such thnt the network connected in series is active to establish connections between the sources and destinations, whi]e the other network is inactive. The arrsngement is such that the means for controlling, upon occurrence of A blocking state in the active network before its maximum capacity is reached and the network is full, includes means for `.
emptying the inactive network of all paths therethrough and then establishing paths through the network which will connect the sources and destinations in the manner attempted when the active network went to a blocking state -- without occurrence of a blocking state in the inactive network. After the paths are established through the inactive network, the selector means is operative to connect the inactive network in series between the sources and destinations and to disconnect the active network from in series, whereby the roles of -the networks are reversed and the previously inactive network becomes active while the previously active -network becomes inactive.
The invention also contemplates a method of n connecting individual ones of a plurality of sources to ~~
individual ones of a plurality of destinations in -~
various combinations. The method comprises the steps of . coupling all of the sources to inlets to each of first -`
and second connecting networks that are nonblocking in .-the wide sense, coupling all of the destinations to 'J'~
outlets from each of the networks, and connecting the first network only in series between the sources and `--destinations. Paths are then established and interrupted through the first network to connect and disconnect individual ones of the sources and ~
destinations. ~pon occurrence of 8 blocking stnte in the first networ~ before it has reached its maximum capacity and is full, paths are established through the second network which will connect the sources and -~
destinations in the manner attempted when the blocking state occurred in the first network -- without causing a blocking state to occur in the second network. The seeond network is then conneeted in series between the sources and destinations and the first network is diseonneeted from in series, whereby the seeond network ~
then eonneets the sources and destinations. ~-The foregoing and other objects, advantages and features of the invention will become apparent upon a consideration of the following detailed deseription, when taken in eonjunetion with the aeeompanying ;~
drawings.
Brief Deseription of the Drawings Fig. 1 is a bloek diagram of a known type of digital switching system, illustrating basie components of the system, whieh inelude a digital eonneeting `~
network;
Fig. 2 is a simplified bloek diagram of a digital , switching system according to the teachings of the present invention;
Fig. 3 is a block diagram of a 4 x 4 connecting network of a type which may represent each network of Fig. 2; and ~ ig. 4 is similar to Fig. 3, except that it illustrates A n x n connecting network. ~
~.~'. .
;..
~ ' ~.144~
Detailed Description The Prior Art Fig. 1 shows in block diagram form a typical'prior art digital switching system ineluding a line and trunk -interface, a digital eonnecting network and a eentra]
control. The line and trunk interface eonneets the system to the outside world, e.g., to serviee eireuits, --i digital trunks, analog trunks and subseriber lines.
.. ...
Where the signals from the outside world are analog, the ~
" .
interfaee eonverts them to digital.
The digital eonneeting network is an arrangement of switehes snd transmission links through whieh terminals connected to sourees of signals may be seleetively coupled in many eombinations to terminals eonneeted to '-destinations, the basie funetion of the digital conneeting network being to perform the neeessary ~--switehing to estsblish the paths for the conneetions.
The eentral eontrol in turn provides stored program ~^~
eontrol for both the syseem and a man/maehine interfaee (not shown). ~--Calls in progress in a digital eonneeting network ~ -~
do not usually arise in a predetermined order or time sequenee. Requests for eonnections (new ealls) and terminations of connections (hangups) occur more or less -~
at random, and unless the network is nonbloeking in the strict sense, or nonbloeking in the wide sense, it can ~-~
happen that the it is not able to establish a connection -~
between a calling source and a selected destination.
~hen that occurs, the network is said to be in a blocking state, i.e., there is an idle inlet to and an id]e out]et irom the network, but there is no availab]e psth ~hrough the network to connect the inlet to the ;:
out]et 7 To AVO id a blocking condition, most new telephone switching s~stems use connecting networks that are nonblocking in the strict sense. However, a strictly nonblocking network is often extensive and complex since nonblocking is normally schieved by an N2 type of network, such that a 2000 port network requires 4 million switch elements to ensure nonblocking.
The Invention -~
~o mitigate cost and eomplexity- disadvantages of conventional switching systems having eonneeting -~
networks that are nonbloeking in the striet sense or :.
nonbloeking in the wide sense, RS seen in Fig. 2 the -~
invention provides an improved switehing system which ineludes a pair of identical conneeting networks 0 and 1, each of whieh is nonblocking in the wide sense.
Every source is connected through each network to every destination, snd elthough only 8 single source and i---destination are illustrated for simplicity, it is -`~
understood that in aetual use, the networks would be conneeted between pluralities of sourees and destinations. -~ -~ . ~
At any given time, only one network aetively establishes paths from sourees to destinations as -~
determined by and through a conventional 1 of 2 selector ;-circuit. The other network is inactive at the time and in a standby mode, although the same paths established -~
through the active network Q150 exist in the inactive ,;
network. This type of redundant configuration of networks increases the reliability of the switching ,~
svstem, such that if the active network fails, the other ~-one will be prepared to maintain the paths and assume switching functions. However, according to the .:.
~t.~44~`
invention, the redundnnt contigurstion also is used to estab]ish an easily implementable nonblocking environment.
The connecting networks O and 1 are n x n three stage networks which are responsive to a central control having a man/machine interfMce (not shown). Connections are added and dropped in a random manner, so blocking can occur. When a blocking state occurs in the active .
network before it has reached its maximum capacity and is full, instead of using a complicated algorithm to .
rearrange existing connections in the active network, -.
according to the ir,vention a nonblocking condition of -the system is easily reached by: (1) eliminating all of the paths in the inactive network, so that the network .
is empty; (2) reestablishing paths in the inactive -~
network, according to a relatively simple algorithm, that will connect the sources to the destinations in the manner attempted when the active network reached R
blocking state, without a blocking state occurring in ~-the inactive network until the network is full; (3) ~-.
switching or reversing the roles of the active and inactive networks, so that the previously inactive _ -network becomes active and establishes the paths from - -the sources to the destinations through the 1 of 2 selector circuit, and the previously active network becomes inactive; and then (4) copying the paths established in the newly active network into the newly ~-~
inactive network, so that the inactive network will be prepared to connect the sources to the destinations and `~
assume switching functions should the active network ~=.-fail. ~
~'~
~ 4, To demonstrate thRt paths may be established throu@h an initially empty n x n three stage network, such thst no b]ocking state occurs unti] the maximum realizable capacity of the network is reached and the network is full, consider the 4 x 4 three stage network shown in Fig. 3, which may represent each of the networks 0 and 1. The network is comprised of square switches, eflch having the same number of inlets AS -~
outlets. ~or a 4 x 4 network, each stage has 4 switches and each switch has 4 inlets and 4 outlets, so inputs Yo-Y3 each represent four lines from respective sources to associated input terminals of source switches So-S3, and outputs Xo-X3 each represent four lines to respective destinations from associated output terminals --of destination switches Do-D3. The network is full àccess, so each switch in a given stage has a transmission link to every switch in adjacent stages, -although for simplicity of illustration only links between the source switch So and intermediate stage ,-switches 10-13, as well as between the destination switch Do and the intermediate stage switches, have been shown.
Starting with an initially empty network, i.e., whichever network 0 or 1 was inactive and emptied of -paths when the other became blocking, to connect the sources and destinations through the inactive network in the manner attempted when the active network reached a blocking state, such that the inactive network does not ~~
become b]ocking until it is completely full, a ~-connection from S~ to Dj is established by choosing an ~-intermediate stage Ij, through which the connection is ~-`
to pass, according to the following algorithm: , -In- ,", ~ 14~
.
Choose the path (Dj, lo), (Dj, Il), (Dj, 12) or (Dj, 13) that has the minimum number of existing connections on that particular link.
After the path is chosen and established, the psth (1;, Sk) is established to complete the connection from Sk to Dj.
The above path choosing and establishing seguence is then repeRted until all of the desired connections that are to exist between sources and destinations have been established. The 1 of 2 selector circuit is then operated to reverse the roles of the active and inactive -networks, such that the previous]y inactive network, with the newly established paths therein, becomes active -and connects the sources to the destinations, and the previously sctive network becomes inactive. The paths in the newly active network are then copied into the newly inactive network, so that it will be prepared to assume connection and switching functions should the sctive network fail.
To appreciate that connections may be Qdded to the ~-emptied insctive network, so thst no blocking occurs until the network is full, considering the following:
Note 1: The capacity of link (Dj, Ij) is not exceeded -~
since each link hHs a maximum capacity of Xmax/4 and Xj/4 < XmaX/4~ where Xj is the traffic leaving Dj and XmaX is the maximum traffic capacity of Dj.
Note 2 Since Xo < Xmax~ , X3 < XmaX~ the maximum traffic through the center stage li is ,~
Xo/9 + X1/4 + X2/4 + X3/4 < 1/4(4XmaX) = Xmax .i-Therefore, the maximum capacity of the center ~-stage is not exceeded.
.
- 1 ~ - ....
~t;144~
~ote 3: Lel tj k = (the number of connections from D
to S~ max~ or the portion of Xj that is connected to Sk, and < tj k < 1. Then, the -~
maximum traffic over link (li, Sk) =
1/4XûtOk + 1/4Xltlk + 1/4X2t2k + 1/4X3t3k , < 1/4Xmax (tOk + t1k + t2k + t3k). ~owever, for a ful] network Xmax(tok + t1k + t2k + t3k) ; = XmaX~ since the maximum traffic entering Sk .i from the left side of the network is also YmaX :i;
= XmaX~ Therefore, for a full network (tOk +
....
tlk ~ t2k + t3k) = 1, although in general, and to include the situation where the network is less than full, (tOk + t1k + t2k + t3k) < 1- . -Thus, the maximum traffic on link (Ij, Sk) -~ -< 1/4 (Xmax). This satisfies the maximum link -requirements of link (1;, Sk).
As seen from note 1, the (Dj, Ij) inter~ediate link capacity is not exceeded; according to note 2, the ~:
center stage capacity is not exceeded; and from note 3 it is seen that the (Ij, Sk) link capacity is not -exceeded. The network therefore remains nonb]ocking when connections are sequentially added to the initially ~-empty network according to the above algorithm.
The proof that paths may be established through an ' `
initially empty three stage network, such that the network remains nonblocking until it is full, may be '~
extended to any n x n network. ~ -Consider the full access n x n network in Fig. 4, `
which may represent each of the networks O and 1.
Starting with the network initiall~ empty, to establish ;~`
B connection from Sk to Dj, choose an intermediate stage ,~
-12- ~
144~, Ij, tllrollgh which the connection is to pRSS, ~ccording to th~ fo~lowing algoritl)m:
Choose the path [(Dj, lo), (D~ ), . . ., or (Dj, In_l)] that has the minimum number of existing connections on that particular link.
After the path is chosen and established, the path (Ij, Sk) is established to complete the connection from Sk to Dj.
The above path choosing and establishing sequence ' -is then repeated until all oi the desired connections that are to exist between sources and destinations have ..
been established. The 1 of 2 selector circuit is then operated to reverse the roles of the active and inactive .:
networks, such that the previously-inactive network, ,'-:
with the newly established paths therein, becomes active ~-and connects the sources to the destinations, and the previously active network becomes inactive. The paths ,-.
in the newly active network are then copied into the newly inactive network, so that it will be prepared to assume connection and switching functions should the i.
active network fail. :-To appreciate that connections may be added to the '~~
network, so that no blocking occurs until the network is --full, consider the following: .
Note 1: The CQpQCity of link (Dj, I;) is not exceeded since each link hss a maximum capacity of max/n and Xj/n < XmaX/n~ where Xj is the ~ -traffic leaving Dj and XmaX iS the maximum ' traffic capacity of Dj.
Note 2: Since Xo < XmQX~ , Xn-l < Xmax~ the maximum traffic through the center stage Ij is ,.
-13- ~
i~il4~2 ~;n/n + Xl/n ~ Xn_l/n < l/n(nXmax) =
max Therefore, the maximum capac;ty of the center stsge is not exceeded.
~ote 3: Let tj k = (the number of connections from Dj to stage Sk)/XmaX~ or the portion of Xj that is connected to Sk, and 0 < tj k < 1. Then, the mnximum traffic over link (Ij, Sk) =
1/n Xoto k + . . . + 1/n Xjt; k + . . +
l/n Xn_1tn_1 < 1/n Xmax (t0,k + + tj,k +
. . . + tn_1 k). However, for a full network Xmax ~t0,k + + tj,k + . . . + tn_1 k) =
XmaX~ since the maximum traffic entering Sk from the left side of the network is also YmaX
= XmaX~ Therefore, for a full network (to k +
+ tj,k + tn_1,k) = 1, a]though in ,.
general, and to include the situation where the network is less than full, (t0,k + + tj,k + + tn-l,k) < 1. ',~
Again, the maximum traffic on link (li, Sk) < -1/n(Xmax). This satisfies the maximum link requirements `-~
of link (1;, Sk). .r "
The invention therefore contemplates a switching .
system having two identical, full access, n x n, three ~-~
stage blocking networks. One network at a given time is : ~.
r active and connects the sources to the destinations, 1.~
while the other is inactive and in a standby mode, but ~-has in it the same paths as are in the active network, ,.
so that shou]d the active network fail, the insctive one ~, will be ready to assume connection and switching functions. The redundancy provides incressed reliability in the event of failure of one of the networks.
Li~l44;2 l`he redundancy also enables the switching system to ~e mRinlained in fl nonb]ocking mode. Although the networks may reach a blocking state, should that occur it may readily be resolved. In particular, upon occurrence of a blocking stnte in the active network, the inactive network is simply cleared of the paths therein, and new paths are established that will connect the sources and destinations in the manner attempted when the active network became blocking, without e blocking state occurring in the inactive network until it is fully loaded. The roles of the active and inactive networks are then reversed.
By virtue of the arrangement, it is not necessary ~`
to interrupt and rearrange connections through the --active network should it reach a blocking state, so there is no interruption of communicfltions services. In addition, the particular- algorithm for establishing the transmission paths through the inactive network is - -~
reletively simple compared with prior algorithms used to esteblish and rearrange paths through networks that are nonblocking in the wide sense. In essence, efter -;
emptying the inactive network, paths are reestablished through its center stage, beginning with lo end ;
sequentially cycling from lo to In_1, to make the ,~
connections between the center stage switches and the various output stage switches Do to Dn_1. After each , path between a switch of the center stage and a switch , of the output stage is established, one is then ~-established between the particular center stage switch -`:
and a selected switch of the input stage, such that a connection from Sk to Dj is completed. The ru]e for ;-msking the connections from S~ to Dj ensures that the . ~
- 1 5 - r::
144~
network m~y ~c completely lorlded ~o its msximum capacity before it renches n blocking state.
~ 'hile embodiments of the invention have been described in detail, various modificntions and other embodiments thereof may be devised by one skilled in the art without departing from the spirit and scope of the invention, as defined in the appended claims.
~.
. . .
.':
...
Claims (16)
1. A switching system for connecting a plurality of sources to a plurality of destinations in various combinations, comprising a pair of connecting networks that are nonblocking in the wide sense, each for being connected in series between all of the sources and all of the destinations; means for controlling said networks to establish and interrupt paths therethrough between individual ones of the sources and destinations; and selector means for connecting one or the other of said networks in series between the sources and destinations, such that the network connected in series is active to establish connections between the sources and destinations while the other network is inactive, said means for controlling, upon occurrence of a blocking state in the active network before its maximum capacity is reached, establishing paths through the inactive network, without occurrence of a blocking state in the network until its maximum capacity is reached, which will connect the sources and destinations in the manner attempted when the active network went to a blocking state, said selector means then being operative to connect the inactive network in series between the sources and destinations and to disconnect the active network from in series, whereby the roles of said networks are reversed and the previously inactive network becomes active while the previously active network becomes inactive.
2. A switching system as in claim 1, wherein said control means is operative, after the roles of said networks are reversed, to establish in the newly inactive network paths identical to those in the newly active network, whereby the newly inactive network is prepared to connect the sources to the destinations should the newly active network fail.
3. A switching system as in claim 1, wherein said connecting networks are identical, three stage, full access, n x n networks.
4. A switching system as in claim 3, wherein each said three stage connecting network comprises an input stage having n x n switches SD - Sn-1, a middle stage having n x n switches I0 - In-1 and an output stage having n x n switches D0 - Dn-1, and wherein said control means, upon occurrence of a blocking state in the active network before the network has reached its maximum capacity, is operative to interrupt any paths in the inactive network and to reestablish paths therein, to connect the sources and destinations in the manner attempted when a blocking state occurred in the active network, by establishing a connection from Sk to Dj by choosing en intermediate state Ij, through which the connection is to pass, by selecting the path [(Dj, I0), (Dj, I0), . . . or (Dj, In-1)], according to which middle stage has the minimum number of connections to output stages, and to then establish a path from l; to Sk to complete the connection from Sk to Dj, said control means then being operative to repeat the establishment of additional paths from Sk to Dj until all of the connections between sources and destinations have been made.
5. A switching system for connecting a plurality of sources to a plurality of destinations in various combinations, comprising first and second connecting networks, each of which is nonblocking in the wide sense and has a plurality of inlets for connection with individual ones of all of the sources, a plurality of outlets for connection with individual ones of all of the destinations, and switch means in circuit between said inlets and outlets and operable to establish and interrupt transmission paths between said inlets and outlets; selector means for selectively connecting either said first or second network in series between the sources and destinations, such that whichever network is in series is active to establish connections between the sources and destinations while the other network is inactive; and control means for operating said switch means of said networks to establish and interrupt transmission paths through the same to connect, through the active network, individual ones of the sources and destinations in a selected manner and, upon occurrence of a blocking state in the active network before its maximum capacity is reached, for operating said switch means of the inactive network to establish transmission paths therethrough, without occurrence of a blocking state in the network until its maximum capacity is reached, which will connect the sources and destinations in the manner attempted when the blocking state occurred in the active network, said selector means then being operative to connect the inactive network in series between the sources and destinations and to disconnect the active network from in series, whereby the roles of said networks are reversed and the previously inactive network becomes active while the previously active network becomes inactive.
6. A switching system as in claim 5, wherein said control means is operative, after the roles of said networks are reversed, to establish in the newly inactive network paths identical to those in the newly active network, whereby the newly inactive network is prepared to connect the sources to the destinations should the newly active network fail.
7. A switching system as in claim 5, wherein said connecting networks are identical, three stage, full access, n x n networks.
8. A switching system as in claim 7, wherein each said three stage connecting network comprises an input stage having n x n switches S0 - Sn-1, a middle stage having n x n switches I0 - In-1, and an output stage having n x n switches D0 - Dn-1, and wherein said control means, upon occurrence of a blocking stage in the active network before the network has reached its maximum capacity, is operative to interrupt any paths in the inactive network and to reestablish paths therein, to connect the sources and destinations in the manner attempted when a blocking state occurred in the active network, by establishing a connection from Sk to Dj by choosing an intermediate stage Ij, through which the connection is to pass, by selecting the path [(Dj, I0), (Dj, I1) . . ..or (Dj, In-1)], according to which middle stage has the minimum number of connections to output stages, and to then establish a path from Ij to Sk to complete the connection from Sk to Dj, said control means then being operative to repeat the establishment of paths from Sk to Dj until all of the connections between sources and destinations have been made.
9. A method of connecting individual ones of a plurality of sources to individual ones of a plurality of destinations in various combinations, comprising the steps of coupling all of the sources to inlets to each of first and second connecting networks that are nonblocking in the wide sense; coupling all of the destinations to outlets from each of the networks;
connecting the first network only in series between the sources and destinations; establishing and interrupting paths through the first network to connect and disconnect individual ones of the sources and destinations; and, upon occurrence of a blocking state in the first network before it has reached its maximum capacity, establishing paths through the second network, without a blocking state occurring in the network until its maximum capacity is reached, which will connect the sources and destinations in the manner attempted when the blocking state occurred in the first network, and then connecting the second network in series between the sources and destinations and disconnecting the first network from in series, whereby the second network then connects the sources and destinations, said establishing and interrupting step then establishing and interrupting paths through the second network.
connecting the first network only in series between the sources and destinations; establishing and interrupting paths through the first network to connect and disconnect individual ones of the sources and destinations; and, upon occurrence of a blocking state in the first network before it has reached its maximum capacity, establishing paths through the second network, without a blocking state occurring in the network until its maximum capacity is reached, which will connect the sources and destinations in the manner attempted when the blocking state occurred in the first network, and then connecting the second network in series between the sources and destinations and disconnecting the first network from in series, whereby the second network then connects the sources and destinations, said establishing and interrupting step then establishing and interrupting paths through the second network.
10. A method as in claim 9, wherein said establishing and interrupting step simultaneously establishes and interrupts the same paths through the second network as through the first, and including the steps, upon failure of the first network, of disconnecting the first network from, and placing the second network in, series between the sources and destinations.
11. A method as in claim 10, including the step, upon occurrence of a blocking state in the first network, and prior to establishing through the second network the paths attempted to be established through the first network at the time it reached a blocking state, of interrupting all paths through the second network so that the network is empty.
12. A method as in claim 9, including the step, after disconnecting the first network from in series between the sources and destinations, of establishing in the first network the same paths as are in the second network, so that should the second network fail, the first network is prepared to be connected in series between the sources and destinations.
13. A method as in claim 9, wherein the first and second networks are identical, three stage, full access n x n networks, each of which has an input stage having n x n switches S0 - Sn-1, a middle stage having n x n switches I0 - In-1 and an output stage having n x n switches D0 - Dn-1, and wherein said step of establishing paths through the second network comprises the steps of interrupting any paths through the network, and then reestablishing paths by establishing a connection from Sk to Dj by choosing an intermediate stage Ii, through which the connection is to pass, by selecting the path [(Dj, I0), (Dj, I1 . . . or (Dj, In-1)], according to which middle stage has the minimum number of connections to output stages, establishing a path from Ii to Sk to complete the connection from Sk to Dj, and then repeating said steps to establish a connection from Sk to Dj until all of the connections between sources and destinations have been made.
14. A method of operating a switching system for connecting individual ones of a plurality of sources to individual ones of a plurality of destinations through one or the other of first and second connecting networks that are nonblocking in the wide sense, comprising the steps of controlling paths through the first network to establish connections between the sources and destinations through the first network only; and, upon occurrence of a blocking state in the first network before its maximum capacity is reached, establishing paths through the second network, without a blocking state occurring in the network until its maximum capacity is reached, that will connect the sources and destinations in the manner attempted when the blocking state occurred in the first network, and then establishing connections between the sources and destinations through the second network only.
15. A method as in claim 14, including the step, while the first network is establishing connections between the sources and destinations, of establishing through the second network paths that will connect sources and destinations in the same manner as they are being connected by the first network, and also including the steps, should the first network fail, of interrupting the connections between sources and destinations through the first network and establishing the connections through the second network.
16. A method as in claim 14, including the step, after establishing connections between the sources and destinations through the second network, of establishing through the first network paths that will connect sources and destinations in the same manner as they are being connected by the second network, and also including the steps, should the second network fail, of interrupting the connections between sources and destinations through the second network and establishing the connections through the first network.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US06/822,389 US4731825A (en) | 1986-01-27 | 1986-01-27 | Nonblocking switching system and method |
| US822,389 | 1986-01-27 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA1261442A true CA1261442A (en) | 1989-09-26 |
Family
ID=25235885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA000522591A Expired CA1261442A (en) | 1986-01-27 | 1986-11-10 | Nonblocking switching system and method |
Country Status (2)
| Country | Link |
|---|---|
| US (1) | US4731825A (en) |
| CA (1) | CA1261442A (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1991014326A2 (en) * | 1990-03-05 | 1991-09-19 | Massachusetts Institute Of Technology | Switching networks with expansive and/or dispersive logical clusters for message routing |
| US5175733A (en) * | 1990-12-27 | 1992-12-29 | Intel Corporation | Adaptive message routing for multi-dimensional networks |
| US5321813A (en) | 1991-05-01 | 1994-06-14 | Teradata Corporation | Reconfigurable, fault tolerant, multistage interconnect network and protocol |
| US5454025A (en) * | 1992-11-12 | 1995-09-26 | Mulrow; Robert J. | Switch bypass for a public safety calling system |
| US6269149B1 (en) * | 1996-12-30 | 2001-07-31 | Paradyne Corporation | System and method for enhancing a communication link |
| US6356622B1 (en) | 1997-05-02 | 2002-03-12 | Paradyne Corporation | System and apparatus for enhancing a network link |
| US6628648B1 (en) | 1998-09-18 | 2003-09-30 | The United States Of America As Represented By The Secretary Of The Navy | Multi-interface point-to-point switching system (MIPPSS) with hot swappable boards |
| US6526048B1 (en) | 1998-09-18 | 2003-02-25 | The United States Of America As Represented By The Secretary Of The Navy | Multi-interface point-to-point switching system (MIPPSS) under unified control |
| US6678268B1 (en) | 1998-09-18 | 2004-01-13 | The United States Of America As Represented By The Secretary Of The Navy | Multi-interface point-to-point switching system (MIPPSS) with rapid fault recovery capability |
| US6426952B1 (en) | 1998-09-18 | 2002-07-30 | The United States Of America As Represented By The Secretary Of The Navy | Multi-interface point-to-point switching system (MIPPSS) having an internal universal signal format |
| US6580720B1 (en) | 1998-09-18 | 2003-06-17 | The United States Of America As Represented By The Secretary Of The Navy | Latency verification system within a multi-interface point-to-point switching system (MIPPSS) |
| US6580692B1 (en) | 1998-09-18 | 2003-06-17 | The United States Of America As Represented By The Secretary Of The Navy | Dynamic switch path verification system within a multi-interface point-to-point switching system (MIPPSS) |
| US6745240B1 (en) | 1999-11-15 | 2004-06-01 | Ncr Corporation | Method and apparatus for configuring massively parallel systems |
| US6412002B1 (en) | 1999-11-15 | 2002-06-25 | Ncr Corporation | Method and apparatus for selecting nodes in configuring massively parallel systems |
| US6418526B1 (en) | 1999-11-15 | 2002-07-09 | Ncr Corporation | Method and apparatus for synchronizing nodes in massively parallel systems |
| US6519697B1 (en) | 1999-11-15 | 2003-02-11 | Ncr Corporation | Method and apparatus for coordinating the configuration of massively parallel systems |
| US6882766B1 (en) | 2001-06-06 | 2005-04-19 | Calient Networks, Inc. | Optical switch fabric with redundancy |
| WO2013028170A1 (en) * | 2011-08-22 | 2013-02-28 | Intel Corporation | Method for data throughput improvement in open core protocol based interconnection networks using dynamically selectable redundant shared link physical paths |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| NL7202501A (en) * | 1972-02-25 | 1973-08-28 | ||
| US4345324A (en) * | 1980-07-09 | 1982-08-17 | Christian Rovsing A/S | Process and system for error detection in a computer-controlled telephone exchange |
| JPS5726955A (en) * | 1980-07-25 | 1982-02-13 | Hitachi Ltd | Backup control system |
| US4497054A (en) * | 1983-04-29 | 1985-01-29 | Honeywell Inc. | One-for-N redundancy in a digital switch matrix |
-
1986
- 1986-01-27 US US06/822,389 patent/US4731825A/en not_active Expired - Fee Related
- 1986-11-10 CA CA000522591A patent/CA1261442A/en not_active Expired
Also Published As
| Publication number | Publication date |
|---|---|
| US4731825A (en) | 1988-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA1261442A (en) | Nonblocking switching system and method | |
| US5331631A (en) | N+K sparing in a telecommunications switching environment | |
| US4656658A (en) | Network routing arrangement | |
| JPH0636619B2 (en) | Reconfigurable multi-connection exchange network | |
| JPH0492319A (en) | Switch module | |
| JP2659421B2 (en) | Self-routing channel | |
| GB2083975A (en) | Back-up control system | |
| JPH0225135A (en) | Method of instructing path to high speed packet switching system | |
| US3582560A (en) | Multistage telephone switching system for different priority users | |
| US5629925A (en) | Method of operating a switching network, as well as switching network and exchange therefor | |
| US5864552A (en) | Rearrangeable non-blocking switching network | |
| CA1296795C (en) | Circuit for a central control time division multiplex pcm telecommunications system with groups of terminals connected to a switching network | |
| IE67126B1 (en) | Method for expanding a normal three-stage switching array | |
| US6240063B1 (en) | 3-staged time-division switch control system | |
| US3068324A (en) | Automatic telecommunication and like switching systems | |
| Lin et al. | On rearrangeability of multirate Clos networks | |
| JP2595804B2 (en) | Exchange equipment | |
| US4023141A (en) | Efficient one-sided rearrangeable multistage switching network | |
| US4750202A (en) | Non-blocking concentrator | |
| US6452929B1 (en) | Network arrangement and method relating to telecommunications | |
| KR19990036761A (en) | Control Architecture for Homogeneous Routing Architecture | |
| EP1106017B1 (en) | Mesh networks | |
| US6324278B1 (en) | Switched telecommunication network, method of establishing a switched connection in such a network, as well as exchange and routing facility therefor | |
| GB2172174A (en) | Switching systems | |
| US20040218590A1 (en) | Telecommunication network with layered switching fabric |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| MKEX | Expiry |