GB1583586A - Switching network - Google Patents

Description

(54) A SWITCHING NETWORK (71) We, INTERNATIONAL BUSI NESS MACHINES CORPORATION, a Corporation organized and existing under the laws of the State of New York in the United States of America, of Armonk, New York 10504, United States of America do hereby declare the invention for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates to a switching network, which can be utilized in digital communication systems.

Communication networks have been proposed which are formed of a plurality of elementary switching units, or exchanges, positioned in geographically distinct locations and connected to one another by digital connections such as Pulse Code Modulation connections, for instance. The configuration of these networks, i.e. the way in which the elementary switching units are connected to one another, is essentially dependent on the geographical location of these elementary switching units and on how high the traffic will be between them.

When considering the high cost of the communications lines, it is obvious that only the geographically closely-positioned elementary switching units will be connected to one another and that a call between remotely-positioned switching units will be set through the intermediary of a plurality of two-by-two-closely-positioned elementary switching units. Likewise, the expansion of this type of switching network is, as a rule, carried out by connecting the newly planted elementary switching units to one of several units comparatively near to one another. It results therefrom very irregular configuration which provides complex paths for routing information from one point to another in the network.

According to the invention there is provided a switching network including a plurality of switching units, each unit being adapted to connect, upon request, lines and/or terminals which are connected thereto to one another, and being defined by a set of three coordinates i,j,k, said coordinates being integers which can assume any of the values from 0 to u for coordinate i, from 0 to v for coordinate j, from 0 to w for coordinate k, and in which said units are connected to one another according to the following rules: a) each unit is connected to those units which have only one coordinate which differs by a unity, the other coordinates being respectively identical, and wherein when i=u, u+l=0, when j=v, j+1=0 and when k=w, k+1=0 and each of u, u and w has value of at least 2, and b) units having all their coordinates but the last one, respectively identical, are connected to one another.

In a system having three coordinates which vary from 0 to u, from 0 to v and from 0 to w, respectively, the set of coordinates i,j,k will be connected to the following sets: According to the first rule: i+1,j,k i,j+1,k i,j,k+1 i-1,j,k i,j-1,k i,j,k-1 According to the second rule: i,j,k+1, i,j,k+2, i,j,k+3, .. i,j,u i,j,k-1, i,j,k-2, i,j,k-3, .. i,j,0 The invention will now be described by way of example with reference to the accompanying drawings in which:: Figure 1 illustrates a switching network embodying the invention; Figure 2 is a schematic diagram of the topographical structure of a network according to this invention; Figures 2a and 2b illustrate a set of elementary units and the connections thereto in the case when the last coordinates can assume four distinct values; Figure 3 is an exemplary illustration of how the connections are set up in the network shown in Figure 2; Figure 4 illustrates the possible routes between three meshes connecting two nodes of the switching network shown in Figure 2; Figure 5 illustrates an example of the data format routed over the digital connections which connect the nodes of the network shown in Figure 2; Figure 6 is a schematic diagram of a procedure for finding out a free route in the network shown in Figure 2; and Figure 7 illustrates all the routes with which it would be possible to embody the procedure shown in Figure 6.

Figure 1 illustrates a switching network including a plurality of elementary switching units each capable of carrying out switching operations in a autonomous way, and comprised of a switching network SW, a control device PROC (for instance, a microprocessor) and memory units MEM for storing, more particularly, the information relative to the condition of the calls in progress; telephone lines of any type (extension lines EXT, trunk lines TK, tie line TL, operators' lines OP, etc.) can be connected to this switching system. These switching operations are carried out either after requests coming from lines catered for by the considered elementary unit of after requests coming from other elementary units and transmitted by the digital switching lines which connect the units two by two.

Each of the elementary switching units is also connected to a common system management device SM which fulfills a number of supervisory functions common to all the switching units, namely, maintenance programs and tests, traffic measurements and other statistical operations, registering and charging of the calls, etc. As this device is no part of this invention, it is mentioned in an illustrative way, only.

As shown in Figure 2, the network is more specifically comprised of twenty-seven elementary switching units distributed into nine identical sets of elementary switching units, each set being itself comprised of three elementary switching units. In other words, the involved system is a system with three coordinates i, j, k wherein each coordinate may assume three distinct values 0, 1 or 2. The first coordinate (i) is indicative of the rank of the set from bottom to top, the second coordinate (j) is indicative of the tank of the set from right to left and the third coordinate (k) is indicative of the tank of a switching unit within a set, clockwise.

According to this invention, the connections between the various elementary switching units in the network are carried out by applying the following two rules: First Rule: Each elementary unit is connected to those elementary units which have only one coordinate different from the others by one unity, the other coordinates being respectively identical, considering that, when a coordinate reaches its highest value, the addition of one unity to this coordinate amounts to give its lowest value. (i.e., in that case, 2+1=0 or 0-1=2).

Second Rule: Those elementary units which have their coordinates respectively equal but the last one (the last coordinate is chosen arbitrarily) are connected two-by-two.

Thus, when applying the first rule to the particular case chosen as an example and shown in Figure 2, there is shown that, for instance, the elementary unit of coordinates 022 must be connected to the elementary units of coordinates 020, 021, 002, 012, 020, 021 (these connections are represented by thick solid lines in Figure 2).

When applying the second rule, there is shown that, for instance, the elementary units of coordinates 020, 021 and 022 are connected two-by-two. The same holds true for the elementary units of coordinates 000, 001, 002 as well as of coordinates 200, 201, 202, etc. The elementary switching units connected two-by-two according to this rule, as a matter of fact, form the threeelementary unit sets, as mentioned above.

It should be noted, here, that when applying the second rule to that particular case shown in Figure 2, this does not define additional connections with respect to those defined by the first rule. This results from the fact that the last coordinate can assume three values, only, that, when adding or subtracting one unity to (from) this coordinate, there are obtained the other two values that this coordinate can assume.

The same would not hold true should the last coordinate assume a number of values higher than three. For a better understanding of that particular case, reference is made to Figures 2a and 2b which illustrate a set of four elementary switching units of coordinates 020, 021, 022 and 023, i.e., a case where the third coordinate can assume four distinct values. Figure 2a illustrates the connections carried out when applying the first rule, and Figure 2b illustrates the connections carried out when applying the second rule. It can be observed that there is superimposition, though not entire as this superimposition was, in the case relative to one coordinate with three values.

As a general rule, the two rules will somehow overlap since, when connecting the elementary switching units in a same set, two-by-two, this necessarily implies con necting each elementary unit in the set to those two elementary units which have their last respective coordinate different by one unity. This overlapping is maximum (i.e., entire) when the last coordinate assumes two or three distinct values, and it is reduced when the number of the possible values that this coordinate can assume, is increased.

There will now be resumed the description of the preferred embodiment of this invention together with the advantages offered by the above-mentioned connection rules.

There can be observed that the adopted connection law leads to a highly regular and homogeneous structure. It will be particularly observed that the rule specifying that 2+1=0 makes it possible to obtain a sort of "looping" the result of which is to reduce to a minimum the number of the meshes to be utilized to go from one elementary unit to another one, and to increase the number of the possible routes.

The thick solid lines of Figure 3 illustrates a possible route to go from elementary unit 021 to elementary unit 202, and to go from elementary unit 010 to elementary unit 221.

It can be observed that each of these routes goes through three meshes, only, and that none of the pairs of elementary units in this system are parted from one another by more than three meshes. It is obvious that, when the traffic in the system is particularly high, a route including more than three meshes may happen to be chosen.

It should also be noted that, the longer the route is (in number of meshes) the higher the number of possible routes with the same length, is. By way of an example, the thick solid lines in Figure 4 illustrate all the possible three-mesh routes (i.e. the shortest routes, indeed) through which it is possible to connect elementary unit 010 to elementary unit 221: these routes are six in number. It is abvious that the number of the routes going through four meshes would be still higher.

These features are extremely favourable since: - only few meshes are necessary to connect two elementary switching units, - the fact of having a high number of short routes makes the finding out of a free route easier even when the traffic is high while the number of busy connections remain limited.

Lastly, it should be noted that the number of the possible routes between two elemen tary switching units depends on: - the number of elementary units involved in one set, - the number of the coordinates assigned to each elementary unit, - the number of the values which can be assumed by each coordinate.

This makes it possible to conceive a flexible system and to adapt it very easily to specific strict needs.

As mentioned above, each elementary switching unit is a small exchange adapted to carry out all the switching and connecting operations between the lines which are connected thereto. However, the switching network according to this invention makes it possible to connect any two lines of the system to each other, i.e., more specifically, two lines which are not catered for by the same elementary switching unit. The connections DL shown in Figures 1 and 2 are more specifically intended to route information between the various elementary units.

These connections are digital connections, for instance PCM connections, i.e., lines which transmit time-multiplex information. Such connections are well known in the telephone field where they are often referred to as "omnibus lines" or "bus".

These omnibus lines carry all the information necessary for the routing of the messages (the term "message" is unilized here in its broadest meaning, i.e., it means the "intelligible information"), and the supervision of the calls (tones, ringing signals and other signalling information), as well as, of course, the messages themselves. All these elements of information are formatted in conventional manner (more particularly in the message transmission technique where the messages are transmitted through meshed networks) by the elementary unit connected to that line where the call comes from. The transmission of this information can be carried out, for instance, according to the point-to-point Binary Synchronous Communications - BSC - technique.

By way of an example, Figure 5 illustrates a format of the type of information transmitted through these connnections. In this format, the first field is indicative of the destination address in the message, i.e., the address of the elementary unit which the called line is connected to, as well as the address specific to this line.

The second field indicative of the address of the line where the call comes from. As previously, this address is formed of the address of the involved elementary unit and of the address of the line, proper.

The third field, which can be comprised of a single bit, is indicative whether the message has to be handled in the transit elementary unit (with a view to reserving a time slot, for instance).

The fourth field is indicative of the type of the transmitted message (voice, data, images, etc.).

The fifth field includes the information utilized to both supervise the transmission of the message and detect the errors.

Finally, the sixth field is the message proper.

Before the information which the message is formed of, is transmitted, it is necessary to determine the path through which this information will be routed from the elementary switching unit where the connection request comes from up to the elementary unit which the called line is connected to. It has been mentioned above that many routes can be utilized to go from one node to another one in the system. (In order to make the understanding easier, the term "node" will be referred to as the point corresponding to the location of an elementary switching unit, in the network shown in Figure 2). With reference to Figures 2 and 6, a simple finding out process will be explained with which it will be possible to find out a free route when only knowing the coordinates of the nodes to be connected.

Before explaining in details the process for finding out a route, it should be kept in mind that each elementary unit is assumed to have at any moment a traffic schedule indicative of the busy condition of the six connections which are associated thereto (i.e., where there is at least one idle time slot in the connection). This schedule is automatically up-dated by the considered elementary unit since it necessarily knows at any moment the condition of its own traffic with the neighbouring nodes.

The problem to be solve is to find out a free route between a node which receives a message (or which sends a message) of coordinates i, j, k, for instance, and a node which the message is to be transmitted to, of coordinates p, q, r.

The general principle of this route finding out process illustrated in Figure 6, is as follows: The respective coordinates of the considered node (i,j,k) are successively compared with those of the destination node (p,q,r) from the left coordinate to the right one.

When there is identity between the two compared coordinates, one proceeds to the examination of the following coordinate.

When there is no identity between the two compared coordinates, if pi1, for instance, reference is made to the traffic schedule to know whether the connection connecting the considered node of coordinates i,j,k to the node of coordinates i+1, j,k has an idle time slot. Should it so happen, connection is made to this new node by reserving a time slot in the considered mesh, and the general finding out process is started anew from this new node. Of course, and this holds true each time connection is made to a new node, the finding out operation is carried out until a new node is reached which has for coordinates three coordinates identical with those of the destination node, namely p=i, q=j and r=k.When the connection leading to node i+1,j,k is not available, the connection which connects node i,j,k to node i-1,j,k is examined in order to check whether it has an idle time slot. When this connection is free, connection is made to this new node by reserving a time slot in the considered mesh, and the general finding out process is started anew from this new node. When this connection is not free, the following coordinate is examined after incrementing counter C by one unity (the part of this counter will be explained further on).

During examination of the second coordinate, everything happens in a way similar to the first coordinate, as shown in Figure 6.

More specifically, it can be seen that when q=j, or when qij but when none of those two links which lead to node i,j+1,k and i,j-1,k is available, the following coordinate is examined (by incrementing counter C in this last hypothesis).

The process utilized for the examination of the last coordinate is, as a rule, similar to the previous ones in so far as, when r is different from k and when one of the connections that lead to nodes i,j,k+1 or i,j,k-1 is available, connection is made to the first of these nodes that is available, and the finding out process is started anew from this new node.

There are, however, modifications when the following hypotheses are made: First hypothesis . r=k. The contents of counter C is checked. When this contents is zero, this means that the other two coordinates are identical with each other, and that the destination node, then, has been reached. The only thing to do is to establish definitely the route found out during the previous steps, by making use of the time slots reserved during the finding out procedure in each mesh.On the other hand, when the contents in counter C is different from zero, this means that pii and/or qij, and that no available connections have been found out to reach nodes i+1,j,k or i-1,j,k or i,j+1,k or i,j-1,k. Therefore, no route is considered as available for the connection of the source node to the destination node, and the finding out process is stopped.

Second hypothesis rik and none of the connections leading to nodes i,j,k+1 and i,j,k-1, are available. Here again, one considers that there is no available route to reach the destination node and the finding out process is stopped. Of course, in the cases when the finding out process is stopped, (be it either completed or not) the counter is reset to zero. The same holds true when, during the examination of the second or third coordinate, the next following node is considered. (When proceeding to the examination of the first coordinate, this resetting operation is not necessary since the counter has not been incremented yet, which means that it is already reset to zero).

The finding out process which has just been disclosed with reference to Figure 6 is but an example applying to the specific configuration of the system, as shown in Figure 2. In this finding out process, it has deliberately been proposed to find out only those available routes which have a predetermined number of meshes (three, in this specific example). By assuming, for instance, that one wishes to connect the node of coordinates 011 to the node of coordinates 212, the finding out process which has just been described makes it possible to find out one out of the eight routes illustrated in Figure 7 by a thick solid line (in case one of them is available).

When making use of an analogous principle, processes can easily be conceived which make it possible to find out other routes in the case when the process which has just been described gives way to congestion. It will suffice, for instance, to proceed with this process allowing, even in case of an identity of the coordinates (q=j, for instance), for a connection with a node having this coordinate which differs by one unity (i,j+1,k or i,j-1,k).

WHAT WE CLAIM IS: 1. A switching network including a plurality of switching units, each unit being adapted to connect, upon request, lines and/or terminals which are connected thereto to one another, and being defined by a set of three coordinates i,j,k, said coordinates being integers which can assume any of the values from 0 to u for coordinate i, from 0 to v for coordinate j, from 0 to w for coordinate k, and in which said units are connected to one another according to the following rules: a) each unit is connected to those units which have only one coordinate which differs by a unity, the other coordinates being respectively identical, and wherein when i=u, u+1=0, when j=v, j+1=0 when k=w, k+1=0 and each of u, u and w has value of at least 2, and b) units having all their coordinates but the last one, respectively identical, are connected to one another.

2. A switching network according to claim 1, in which said last coordinate can assume at least three distinct values.

3. A switching network according to claim 2, in which digital connections are provided by lines connecting said switching units to one another.

4. A switching network substantially as hereinbefore described with reference to the accompanying drawings.

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (4)

**WARNING** start of CLMS field may overlap end of DESC **. is considered. (When proceeding to the examination of the first coordinate, this resetting operation is not necessary since the counter has not been incremented yet, which means that it is already reset to zero). The finding out process which has just been disclosed with reference to Figure 6 is but an example applying to the specific configuration of the system, as shown in Figure 2. In this finding out process, it has deliberately been proposed to find out only those available routes which have a predetermined number of meshes (three, in this specific example). By assuming, for instance, that one wishes to connect the node of coordinates 011 to the node of coordinates 212, the finding out process which has just been described makes it possible to find out one out of the eight routes illustrated in Figure 7 by a thick solid line (in case one of them is available). When making use of an analogous principle, processes can easily be conceived which make it possible to find out other routes in the case when the process which has just been described gives way to congestion. It will suffice, for instance, to proceed with this process allowing, even in case of an identity of the coordinates (q=j, for instance), for a connection with a node having this coordinate which differs by one unity (i,j+1,k or i,j-1,k). WHAT WE CLAIM IS:

1. A switching network including a plurality of switching units, each unit being adapted to connect, upon request, lines and/or terminals which are connected thereto to one another, and being defined by a set of three coordinates i,j,k, said coordinates being integers which can assume any of the values from 0 to u for coordinate i, from 0 to v for coordinate j, from 0 to w for coordinate k, and in which said units are connected to one another according to the following rules: a) each unit is connected to those units which have only one coordinate which differs by a unity, the other coordinates being respectively identical, and wherein when i=u, u+1=0, when j=v, j+1=0 when k=w, k+1=0 and each of u, u and w has value of at least 2, and b) units having all their coordinates but the last one, respectively identical, are connected to one another.

2. A switching network according to claim 1, in which said last coordinate can assume at least three distinct values.

3. A switching network according to claim 2, in which digital connections are provided by lines connecting said switching units to one another.

4. A switching network substantially as hereinbefore described with reference to the accompanying drawings.