EP1108304A1

EP1108304A1 - Method for routing messages between two access points

Info

Publication number: EP1108304A1
Application number: EP00948056A
Authority: EP
Inventors: Pierre Dupuy
Original assignee: Alcatel CIT SA; Alcatel SA
Current assignee: Alcatel CIT SA; Alcatel Lucent SAS
Priority date: 1999-06-24
Filing date: 2000-06-22
Publication date: 2001-06-20
Also published as: FR2795593B1; FR2795593A1; AU6164300A; CN1315088A; WO2001001616A1

Abstract

The invention concerns a method for solving the problem of under-utilization of channels (1-2) set up between two access points (A -B) which consists in monitoring between said two access points the number of channels in service and the utilization rate of said channels. When the global utilization rate of said channels turns out to be less than a nominal capacity minus one of the number of said channels, said channels are re-organised (5) so as to utilize them as closely as possible to their nominal capacities. Thus, the method enables to reduce the number (N) of equipment items to be installed in a routing node (A) in a network, while maintaining the same nominal performances or for a given number of equipment items to increase the processing rate of the node.

Description

Method for routing messages between access points

The present invention relates to a method of routing messages between access points. It is more particularly used in the field of telephony for the production and improvement of switching exchanges. It is in particular usable, in an improvement, in the field of mobile telephony. The object of the invention is to reduce the cost of telephone exchanges.

In the field of telephony, it is customary to arrange switching systems between so-called circuit switching systems and so-called channel switching systems. Circuit switching systems amount to establishing, between two access points to be connected, a physical circuit, a pair of copper wires, a fiber optic link with all the necessary equipment or another permanent link. This permanent link is left entirely available to subscribers located on either side of the two access points to be linked. Under these conditions, they can exchange information on this line at their disposal with as high a speed as they wish, to the extent that, technologically speaking, this available link allows it.

In contrast, in channel switched systems, a network between multiple access points is organized and the connection between one subscriber and another via this network takes, at a given time, a channel in a channel of this network. A channel can thus be defined as a reservation of a time window of a frame, as an allocation of a frequency bandwidth in a global band, or both simultaneously. It can also be defined as a temporary assignment of a coding law in a coded division multiple access (CDMA) multiple access channel. Under these conditions, the essential difference between circuit switched systems and channel switched systems lies in the fact that the message sent in the second case includes the address of the subscriber receiving the information sent. In fact, in circuit switched systems once the connection is established, it is not necessary to introduce an address in the message to be transmitted since this can only be transmitted from one point to another between the connected points.

In contrast, in a channel switching system, routers are installed at each node in the network. Each time these routers receive a fraction of a message to be transmitted, they extract from this message the address part and assigns to this message, to transmit it further on (generally towards the subscriber), a channel which connects this router to another router to transmit a packet of information contained in this message. Then the message with its information packets progresses in the network and is finally transmitted on the line of the recipient subscriber.

The second system has the advantage over the first that the infrastructure to be put in place is much less important. On the other hand, this results in a permanent need, for each packet of information to be transmitted from one subscriber to another, to extract the address of the recipient, to determine a path by which this packet of information must be transmitted further and insert this information packet into a channel of a channel corresponding to this path. The determination of the path involves the designation of time windows, bandwidths, coding keys and, most of the time, destination routers in which the rest of the routing processing will be carried out. The conditions of use of the network, the cessation of a communication between two subscribers and the establishment of a communication between two new subscribers lead to the fact that the routing of messages between the connected subscribers is constantly changing, constantly reorganized . The invention relates to this reorganization.

The realization of the routers requires the installation in the access points, one also says the nodes, of electronic circuits of commutation and transmission capable of constituting on demand channels and channels of transmission in sufficient number for communications to route. The problem of producing these electronic circuits is that of optimizing the hardware means used, of the minimum number of these electronic circuits necessary for a nominal number of communications to be routed. The most important problem to be solved is that of the optimized use of the channels. One can, by simplifying in a case which will be retained elsewhere in this description, to say that a channel in a router is constituted by a certain number of time windows, for example L = 32 time windows of a frame. In other words, such a channel is capable of allowing the transmission of L = 32 different messages simultaneously. This channel then has 32 channels. The circuits of such a router include time multiplexing circuits in insertion and time demultiplexing circuits in extraction. During a session, such a channel is used for communication between an access point A and an access point B. This channel here requires the presence of an insertion multiplexer capable of receiving L = 32 messages.

One can imagine that this time insertion multiplexer is produced on an electronic card. If, for statistical reasons at a given moment, there is only one message left which must be exchanged between a router A and a router B, the channel must nevertheless be maintained. The electronic card concerned only serves to carry a single time window, a single channel. In practice, the other time windows are filled with jamming signals. In digital transmission, these stuffing signals are 1s or 0s, or channel coherence signals. If L - 1 = 31 other subscribers then want to establish communications between them and other accessible subscribers, no longer by access point B, but by other access points C or D, the established channel cannot not be used. In this case, you have to use other electronic equipment, other electronic cards, to set up other channels.

Document US-A-4,718,058 also discloses a channel switching system in which messages of reduced size are concatenated together to occupy an entire channel in a channel, a virtual link between two access points. This approach, which is ultimately similar to the establishment of channels in a channel between two access points tends to solve the problem of non-optimal use of the channels. However, it suffers from the same drawbacks as those mentioned above. The expected material gain, which could be for example of the order of four insofar as four messages are concatenated, with a speed at 8 Kbits / s to form a global message at 32 Kbits / s (and which would use a channel of channel), is in fact illusory. In practice, the material gain observed rather than being a 4 to 1 ratio is only a 1.3 to 1 ratio. Indeed, the statistical phenomena of occupation and the diversity of occupancy requests make the channels underused, and therefore require significant use of tamping.

Similarly in document WO-97-27720, provision is made for adding an auxiliary switching circuit which makes it possible to constitute messages with a higher bit rate, for example 32 Kbits / s from messages obtained with a lower bit rate by example 16 Kbits / s or 8 Kbits / s. In theory, as in the previous case, this system works well. In practice the gain in material is only 30% on the realization of the multiplexers. It may be more important. However, this greater gain is achieved at the expense of the considerable increase in the size of an auxiliary circuit. This leads to a misleading assessment of the reduction in the price of the telephone exchange.

In the invention, to remedy this problem, and in particular the problem of underemployment of the channels established between two access points, provision is made to monitor the constitution of these channels. When it thus appears that between two access points a number of channels is abnormally large for a global number of channels provided between these two access points, it is decided to reorganize these channels so as to create full channels, with each if possible a nominal number of channels. In this way we free up channels. This allows in practice to release electronic circuits, electronic cards with which these channels have been established. Under these conditions, these electronic circuits can then be used to establish other communication channels, in particular with other access points.

We will show that by doing so the gain goes from 1, 30 to 1 to 3 or 4 to 1, on average to 3.5 to 1. The example treated will be that of the constitution of channels at 32 Kbits / s from 8 kbit / s outgoing channels. The theoretical limits are almost approached with the invention. To simplify, we can say that the reorganization carried out in the invention consists in moving an allocation, devolved to a message, from a time window of a channel to a time window, for example of the same rank, in another channel, in order to free a channel and therefore the electronic circuits which correspond to it. In these conditions this freed channel can be used to create communications with other access points. In practice if synchronization problems arise at the time when this routing switching must be carried out, it will be preferable in the field of mobile telephony to cause a base station transfer for a mobile telephone whose communication must be routed by a other channel. By doing so, we will show that we do not lose any information from those that were to be transmitted, that we do not send the same information packets twice, and that we do not make an additional auxiliary circuit.

The subject of the invention is therefore a method of routing messages between access points in which

- a framing is constituted and a group by time multiplexing of messages originating from various origins according to frames of this framing,

channels are established between a first and a second access point, a channel thus established having a nominal capacity in number of messages,

- frames of this screening are sent from the first access point to the second access point using the channels set up between these access points, characterized in that

- we measure the number of messages sent between this first and this second access point, and we measure the number of channels set up between this first and this second access point to send these messages, - we reorganize the use of the channels established between this first and this second access point when the number of messages sent by the channels established between this first and this second access point is less than a nominal capacity of all of these instituted channels minus one.

The invention will be better understood on reading the description which follows and on examining the figures which accompany it. These are presented for information only and in no way limit the invention. The figures show:

- Figure 1: a schematic representation according to the invention of a routing circuit, or router, mounted in an access point, receiving messages and which is intended to transmit them to other access points; - Figure 2: a representation of a screening created to allow an exchange between access points;

- Figure 3: the representation of the material means necessary to set up in the access points, and of the channels that these material means make it possible to institute between different routers. Figure 1 shows a system usable for implementing the message routing method of the invention. It essentially shows an access point A. The access point in practice is geographically located at a place in a territory. For example, it is a telephone exchange in a small town or a district of a large town, or possibly a relay center installed at a routing node of a network. Access point A is equipped with circuits to allow it to form channels with other access points, for example here access points B, C or D.

The circuits available for this purpose in the routing circuits at the access point A will in one example include N time multiplexers. In an example to simplify N will be 10. These multiplexer circuits include N combination circuits such as 1 connected at input to N delay circuits 2. Delay circuits 2 have L inputs (L = 32 in an example) and L outputs connected to L inputs of each of the N combination circuits such as 1. The delay circuits 2 each comprise L delay lines such as 3, programmable, making it possible to constitute, in a given framing, frames of L messages. The N combination circuits such as 1 each have an output 4 connected to an input of a switching circuit 5 which makes it possible to connect the output 4 to a line, for example physical, 6 to 8. The lines 6 to 8 allow d '' forward messages from access point A to access points B or C or D respectively. Circuit 5 makes it possible to make a choice between these lines, and therefore to choose a destination. The symbolic representation of Figure 1 can take various material forms. Possibly the physical lines 6 to 8 may themselves be only channel channels with an even higher speed.

Corresponding equipment is created in access points B, C and D. It will already be noted that, for a number N of multiplexers, the possible number of access points B, C or D should itself be limited to N At least in practice, because if this is not the case, it is possible that an access point is the recipient of a communication and that no channel can be allocated to this call given that all the channels would have been allocated to the other destinations. To ensure the functioning of the routing circuits of the access point A, a processing processor 9 issues orders O to the circuits 1, 2 and 5 and is connected for this purpose by a bus 10 to these circuits, to a program memory 11, to a data memory 12 and to a clock 13. The program memory 11 comprises a program 14 which, in a known manner, comprises a multiplexing subroutine 15.

Figure 2 shows the constitution of a screening. For example, in constituted screening, frames of 125 μs are each made up of 32 time windows. A timing at 8000 Hz of circuits 1, 2 and 5 synchronized with the signal from the clock H allows in this case to route L = 32 messages from various origins.

For example, these messages come from mobile telephones 16. Here an uplink is shown, from a mobile telephone 16 to a base station 17 connected itself, downstream in the network, to the access point (to the network node ) A. For example, the mobile telephones 16 comprise, for this uplink, in the case where the message is a speech message, a microphone 18 connected to an analog-digital sampler-converter 19. The sampler-converter 19, of in a conventional manner, is clocked by a signal at 8 kHz without there being any a priori parallel between this sampling rate of the speech signal and the rate imposed on the routing circuit of the access point A. D ' in a known manner, the sampler-converter 19 produces 13-bit words per sample. These 13-bit words are compressed into 8-bit words in logarithmic compressors 20 called Law A type. The 8-bit words are then applied to the input of a transceiver 21 connected to a transmission aerial. The 8-bit words are thus successively sent by the air to a base station 17 with which the mobile telephone 16 is in contact. It is thus conceivable that L mobile telephones such as mobile telephone 16 are connected to a delay circuit such as 2 which is connected to combination circuit 1. In this case, we are able to constitute an entire frame to be routed between the circuit 1 and an access point, access point B for example. In practice, switching circuits take place between circuits 1 and 2. FIG. 3 shows, for the N multiplexers of FIG. 1 located at the access point A, the possibility of creating, for each, channels 6 or 7 or 8 leading to the access points B, C or D respectively. It is understood that from the moment when a multiplexer 1 - 2 has been allocated to a given link, for example the link A - B, throughout the duration of frame 22, the network is in a situation equivalent to a network with circuit switching. It can thus be assumed that at most N multiplexers make N connections between the access point A and the access point B at least in theory. In practice, each of the N multiplexers of A is capable of establishing a link with the access point B.

We can now give a numerical explanation (arbitrarily simplified) of the problem solved by the invention and as it really presents itself. If N x L (i.e. 10 x 32 = 320) communications are to be established between A and B, the N multiplexers of node A are devoted to a link between A and B. If then N x (L - 1) communications cease and that however a communication persists occupying a channel in each channel between A and B then the N multiplexers continue to be all occupied (to route each a single message!). All the resources of access point A are then used. Under these conditions, a next communication which must be created cannot be carried out if the corresponding message must be routed on an access point C. On the other hand, if this message must be routed on the access point B this message will be naturally incorporated in one of the channels already established between A and B. But as statistically the request to the access point C is possible this amounts to having to create additional multiplexers. The total of the multiplexers (and electronic cards) then becomes a greater number than N to equip the access point A. These additional multiplexers are essential although this access point A already has a nominal capacity of N x L communications. We can show in practice that the number of multiplexers is of the order of N x L x 3 when we would like it to be of the order of N only.

In the invention, the channels established between the first access point A and the second access point B are monitored. And for all the channels thus identified, the number of messages, the number of channels actually used in each of these is measured. canals. To this end by example, but shown here for explanatory purposes only, in memory 12, a record of a database can be assigned to each of the institutable channels, comprising a channel name on the one hand and a channel name on the other hand, and L additional fields corresponding to each record in this database an indication 24 mentioning the occupation of a corresponding channel in this channel.

In memory 12, there are as many databases as there are links between different possible access points. For example in this case there will be three, one for the connection between A and B, one for the connection between A and C and one for the connection between A and D. In practice this monitoring is not complicated to implement since the processor 9 allocates, by virtue of the multiplexing program 14, the different time windows of the channels to the messages to be transmitted. Therefore memory 12 and the databases it contains are already available in the state of the art, at least in an implicit manner.

A monitoring program 25 according to the invention therefore makes it possible to identify for a link between two access points, for example A and B, the number of occupied time windows as well as the number of channels formed among the N different multiplexers. This is preferably done in each of the links, AB, AC, or AD. Then we look for a number of communications thus effectively exchanged if this number is less than a nominal capacity of all the channels instituted minus one. Ultimately, in the case where all the multiplexers are capable of M communications (which a priori is not an obligation), we see if the number of communications for M channels is not less than L x (M - 1 ). If so, this means that at least the equivalent of one channel can be saved by rearranging the channels in those channels. In this case, the invention is reorganized by applying the program 25. For this purpose, it is preferably sought which is the least occupied channel. Let M be this channel. The reorganization comprises in principle that a channel using a time window 26 of the channel of the multiplexer M is transferred in another channel, for example the channel of the multiplexer 1, to a time window 27 posterior (or anterior to the time window 24) . Preferably window 27 has in this channel of multiplexer 1 the same time position as that which it occupied in channel M. So step by step you can transfer the communications routed by one channel (and therefore by a given multiplexer) to another channel. Preferably, the channel that we decide to release will be the channel that will have the fewest messages transmitted. The reception channels are preferably the busiest channels among the M - 1 remaining channels.

However, it is possible that the time position of a message in a frame does not meet a free equivalent window in the other available channels where it is likely to be transferred. In this case a problem arises during the transfer. Without going into details, we know that this type of situation commonly leads either to provide the same information twice, or not to send the content of a message corresponding to a time window. In the case of voice message transmissions, it could be accepted that such transmission errors are present. In this case, only the content of a voice message corresponding to 125 μs would be altered in the example chosen. Although unsatisfactory, such degraded operation could be accepted. In this case, rather than transferring the time window 26 to a time window 27, it can be transferred to a time window 28, in another channel, but which is situated on another date relative to the synchronization of the start of the frame. .

However, in the case where binary data to be transmitted is not speech data but pure digital data, such a solution is not admissible. As in the access point A we are not aware of the nature of the information transmitted, we will prefer in the invention to act differently. The goal remains to constitute transmission channels 1 to M - 1 as compact as possible so as to put the least possible filling data and so as to save as much resources as possible. If this problem is to be solved, it is then necessary to carry out heavy management of the transmitted information packets at the access points A. In fact, the delay circuits 3 of the multiplexers 1 - 2 are ultimately only capable of delaying a signal for a duration equal, at most, to the duration of a frame. However, in the case where a transfer would cause the content of a time window 26 to be lost (because it should be placed in an earlier time window 28 in another channel), the content of time window 26 should be stored. , at moment of transfer, for a duration greater than the duration of the frame.

Solving the problem in this way in fact amounts to making the multiplexing circuits 1 very complex and therefore making them lose the interest of simplicity their structure naturally allows. On the other hand, in the case where the time window 26 has to be placed in another channel at a later position 28, the problem does not arise.

In the invention, in the context of mobile telephony, this problem is very simply solved without having to add additional equipment, without having to supplement the delay circuits 3 with delay lines of very long duration.

We have in fact realized that in the context of mobile telephony, in particular of the GSM type, at the time of so-called handover operations, which correspond to the transfer of a link between a mobile telephone 16 and a base station 17 at a link between this mobile telephone 16 and another base station 29, this problem of lost time window was already managed. In short, in the base stations 17 and 29 there already exist all the equipment necessary so that the subscriber does not even realize that he has gone from his relationship with a base station to a relationship with another base station. This transfer mode is particularly rigorous in data transmission.

In the invention, when we want to move a message from a time window 26 in a free channel to another time window, we cause a transfer from base station to the communication exchanged between the telephone 16 and the base station 17. Under these conditions, channel 26 appears, seen from node A, as a channel which ceases to exist in favor of another channel 28 created. This new channel created 28 is of course, in a known manner, assigned to one of the channels already existing between the access point A and the access point B. In the invention, among the M candidate channels to receive this channel , the channel M is eliminated in order to retain only M - 1. In the invention, on the one hand, the least busy channel, here the channel M, is sought, and then the reuse of this channel is prohibited as long as the M - 1 other channels established are not all complete. If necessary, this ban can be lifted before full use of the M - 1 other channels. In this case, the control circuits of the access point A send by means of a hand-over program 30, HO, to the base stations 17 and / or 29, or possibly to a general management circuit for these base stations, a signal S constituting a transfer order. More precisely, the signal S will appear as a transfer request. This request does not need to be satisfied immediately. It can be satisfied one or more frames later. What matters is that overall for a long period, in particular that of the conversations exchanged with mobile phones, the data specific to these communications are routed on a compacted channel and not on very diverse and underemployed channels. If necessary, the occupation of another channel during the same frame 22 for another channel but for the same communication can occur before the release of the channel M.

Claims

1 - Method for routing messages between access points (A - D) in which - a screening (22) is formed and a group by time multiplexing (1 - 2) of messages from various origins (16) according to a frame of this screening,

- channels (Mux1) are established between a first (A) and a second (B) access point, a channel thus established having a nominal capacity (L) in number of messages,

- frames of this screening are sent from the first access point to the second access point using the channels set up between these access points, characterized in that - the number of messages sent between this first is measured (25) and this second access point, and the number (M) of channels established between this first and this second access point is measured to send these messages,

- the use of the channels set up between this first and this second access point is reorganized (30) when the number of messages sent by the channels set up between this first and this second access point is less than a nominal capacity (L x (M - 1)) of all of these established channels minus one.

2 - Method according to claim 1, characterized in that

- it is implemented in a mobile telephone network (16, 17, 29), and

- To reorganize, one causes (5) a transfer of base station to communications exchanged by mobile telephones transmitting messages whose routing undergoes this reorganization.