FR2711468A1 - Interconnection device between two remote local networks, and corresponding interconnection method. - Google Patents

Abstract

The invention relates to a device for interconnecting at least two remote stations and / or local networks by means of at least two communication networks, comprising on the one hand at least a first packet communication network and 'on the other hand at least one second circuit-switched communication network, said device comprising means (25) for recognizing the type of an application flow to be transmitted, and means (22, 28) for automatically selecting one of said communication networks for said communication to be established, as a function of said type of flow. In an advantageous embodiment, said selection means (28) also take into account the communications already established (29) between the same stations and / or local networks, so as to use the same communication resource if possible. Thus, the communication costs are greatly reduced.

Description

Interconnection device between two distant local networks, and

  corresponding interconnection method.

  The field of the invention is that of digital data exchanges between local networks and / or remote data processing stations, via

communication networks.

  More specifically, the invention relates to an interconnection device between two

  elements (networks and / or stations) distant, and a corresponding interconnection method.

  Data transfer operations via a communication network are often referred to as the routing term. To do this, a local network is associated with an interconnection device, or router, capable of providing a connection to a network

  external to exchange data with a remote element, via this external network.

  Two major types of external communication networks can be distinguished, namely packet communication networks (for example according to the X25 standard), and more generally by protocol data units, and circuit-switched communication networks, ensuring a permanent connection for the duration of a

  communication (eg an integrated services digital network, or ISDN).

  Such communication networks are managed by operators, public or private, who of course charge the use of their network. This billing takes into account many criteria, depending on the type of network. In the case of a packet network, the main criterion is usually the amount of data exchanged. In the case of a network

  ISDN, on the other hand, the length of the call is essentially taken into account.

  In general, it is desirable to limit the costs of communications.

  However, this is not easy because the respective costs of communication on one or other of the types of external networks available vary widely depending on the type of the data flow exchanged. For example, if the application flow corresponds to a file to be transmitted in a single operation, a communication billed only to the duration (on ISDN for example) is preferable. On the other hand, if the flow corresponds to an interactive application (consultation of a database, for example), it is desirable that the communication be invoiced on the quantity of information exchanged,

  whatever the duration of the communication.

  Some routers allow the operator to choose the communication network he wishes to use for a particular data exchange. However, they are complex to use because each communication requires specific programming, depending on the network chosen, and requires the operator to be able to determine what will be

  the least expensive network, which is not always obvious.

  The invention aims in particular to overcome these disadvantages of the state of the

technical.

  More specifically, an object of the invention is to provide an interconnection device, or router, of simple implementation and whose cost of use is

  limited, compared to existing routers.

  Another objective of the invention is to provide such a device allowing the transmission of several different application streams, while limiting the costs of use.

external networks.

  These objectives, as well as others which will appear subsequently, are achieved according to the invention by means of a device for interconnecting at least two stations and / or local area networks via least two communication networks, comprising on the one hand at least one first packet communication network and on the other hand at least one second circuit-switched communication network, comprising means for recognizing the type of an application flow at transmit, and automatic selection means of one of said communication networks for said

  communication to establish, depending on the type of flow.

  Thus, the device of the invention, for example a router, optimizes data exchanges automatically. In other words, the device provides a routing that adapts, depending on the type of application flow to be transmitted (for example a file or a file).

interactive application).

  Advantageously, the optimization takes into account at least one of the criteria belonging to the group comprising: the monodirectional or bidirectional aspect of the communication; - the continuous or episodic aspect of the communication; - the response time of said communication networks - the duration of the communication; - the speed of communication; - available communication resources; the billing costs and conditions associated with each of said communication networks. According to the invention, the first packet-switched network may be an X network 25, such as, for example, the TRANSPAC (registered trademark) network. The second circuit-switched network is for example an ISDN network, of which one or more channels are used. In another embodiment of the invention, it is also possible to use an integrated services digital network (ISDN) comprising at least one digital information transfer channel (B) and at least one signaling and transfer channel. packets (D). In this case, said selection means ensure the selection of at least one of said channels as a function of said type of application flow, said D channel.

  corresponding to said first network and said B-channel corresponding to said second network.

  In this embodiment, it should be noted that the term network within the meaning of the present patent application is different from that designating an ISDN network. It must be taken in the more general sense of communication medium (designating in particular a

  network (ISDN or X25), or a channel of an ISDN network, etc ...).

  According to an important characteristic of the invention, said selection means also take account of communications already established between the same stations and / or

  local networks, so as to use the same communication resource if possible.

  It is thus possible to greatly reduce the costs of data exchanges.

  Furthermore, always with the aim of limiting costs, the device of the invention advantageously comprises data compression means associated with at least one of said communication networks, and in particular means for compressing data.

  data transmitted on said first network.

  The invention also relates to a method for interconnecting at least two remote stations and / or local area networks via at least two communication networks, comprising on the one hand at least one first packet communication network and secondly at least a second circuit-switched communication network, comprising the following steps: recognition of the type of application flow to be exchanged; automatic selection of one of said communication networks for said

  application flow, depending on the type of flow.

  Advantageously, said automatic selection step comprises, for at least certain types of flow, a hierarchical selection procedure among at least two

  types of communication networks, depending on the networks already used.

  In this way, the selection of a communication network is adaptive, and takes into account other application flows to the same recipient, in particular to use,

  where possible, a single line of communication.

  The method of the invention therefore advantageously comprises a communication optimization step, aiming to circulate at least two application flows.

  of different types on the same line of communication.

  In a particular embodiment of the invention, said types of application flows include: file flows, corresponding to file transfers; - interactive flows, corresponding to interactive data exchanges; and

  - the default flows corresponding to the other flows.

  In this case, the optimization comprises the following steps: - if the stream is a file stream and no path to the recipient is open, opening a path on a permanently connected communication network for the duration of one communication; if the stream is an interactive stream and no path to the recipient is open, opening a path on a packet communication network; - if the stream is a file stream and a path to the recipient is open on a packet communication network, opening a path on a circuit switched communication network and closing the open path on a communication network in packets; - if the feed is an interactive feed and a path to the consignee is open

  on a circuit switched communication network, maintaining said path.

  Advantageously, when a path is open for the transfer of at least two distinct application flows and an application flow is completed, the selection operation

  an optimal communication network is reiterated.

  In a preferred embodiment, the method comprises a programming step for associating with at least one of said types of application flows and / or

  at least one recipient at least one preferential communication network.

  Other features and advantages of the invention will appear on reading the

  following description of a particular embodiment of the invention, given as a

  illustrative and non-limiting example, and accompanying drawings in which: - Figure 1 illustrates an exemplary system comprising a plurality of local networks interconnected via two separate communication networks using routers according to the invention; FIG. 2 is a simplified block diagram of a router of FIG. 1; FIG. 3 presents a block diagram of the connection method of the router of

Figure 2.

  The device of the invention is intended to interconnect several local networks, or independent processing stations. Figure 1 is an example of a system of

  communication implementing such devices, hereafter called routers.

  This system comprises three adaptive routers 111, 112 and 113, which can communicate via two external communication networks: a packet-switched network 121, for example in accordance with the X25 standard;

  a circuit switched network 122, for example of the ISDN type.

  It allows the interconnection of local networks 131 to 134, as well as

  independent processing 141 or 142.

  The routers according to the invention are preferentially connected to the two communication networks 121 and 122 (routers 1 and 113). However, if necessary, they can of course be connected to a single network (router 112). A router (111) may optionally provide outward links for several local area networks 131 and 132. Two situations may for example be envisaged for packet communication networks 121: it may be an independent X25 network

  (Eg TRANSPAC) or D-channel of an ISDN system.

  Figure 2 illustrates the general principle of a router 11 i according to the invention. The application flow 21 to be transmitted is directed to a selector 22 which directs it to either a network

  packet communication 23, or to a circuit-switched network 24.

  Furthermore, the application flow 21 is received by a recognition module 25, which determines the type 28 of this application flow. Based on this type 26 information, a

  driver decision module 28 the selector 22.

  The recognition operation consists for example in extracting from a datagram information describing the type of flow. For example, three types of application flows can be provided: the "file" type, which in principle is preferable to transmit on a circuit-switched network; - the "interactive" type, which in principle is preferable to transmit on a packet-switched network; - the "default" type, for all applications that do not belong to one of the

two previous types.

  To route according to the type of application, the administrator must first define his applications. Then, it must define the routing to each possible destination. The definition of the routing to a given destination (machine or network) passes in this case by the definition of a set of paths, for which we assign a selection order

by type of application.

  Of course, it is possible to inhibit the adaptive routing function according to the invention to a destination. In this case, the different paths will be used indifferently for all applications. In addition, the administrator can prevent an application from being executed to a given destination, for example to limit the costs of

  communication (interactive to an off-border destination, for example).

  Moreover, the invention ensures the optimization of costs in case of mixed application flows. Indeed, the decision module 28 does not only take into account the flow type information 26, but also the other flows 29 being transferred.

to the same recipient.

  For example, if we designate the different flows as follows: file flow: flow corresponding to file transfer type applications; - interactive flow: flow corresponding to interactive type applications; - default flow: flow corresponding to the applications belonging to neither of the two preceding types

  makes the applications not declared by the user).

  It will generally be considered that the default flow is closer to the interactive flow than to the file transfer flow, since if the user does not explicitly declare a

  application is that it considers as secondary its need bandwidth.

  When a path is opened for the "file" stream, the router also uses it to route the "interactive" or "default" streams until the end of the file transfer. the router then switches the flow "interactive" or "default" on the

  path best suited to these types of flows.

  On the other hand, if a path is already open for the "interactive" or "default" flow, the router does not reuse the latter to convey a "file" flow, instead it opens a new path for the file transfer, and toggles the streams "interactive" or "default" on

this last.

  The purpose of this optimization is to reduce billing by considering that if a big "pipe" is open for file transfer, we can take advantage of it to pass additional sporadic information. This optimization nevertheless respects the choices of the user anxious to use

  the widest possible bandwidth for file transfers.

  Figure 3 is a flowchart illustrating this technique in more detail.

  The first step 32 consists of analyzing the type 31 of streams to be transmitted. If it is a flow 33 "'T'" (interactive) or "D" (default), it is checked (34) if an X25 communication is already open. If the answer is affirmative (35), the new flow is transmitted (36) on the

same communication.

  If the answer is negative (37), check (38) whether a B-channel is open. If so (39), this channel B is transmitted (310). Thus, while an interactive flow is in principle preferably transmitted over an X25 communication, the adaptive router chooses

  automatically channel B already established to limit costs.

  In the opposite case (311), opening (312) an X25 communication, then

  transmits (313) on this open communication.

  If the stream to be transmitted is of type "F" (file) 314, the router looks (315) if an X25 communication is open. If this is the case (316), this communication is closed (317), a channel B is set up (318) and the stream "file" and the stream that was previously transmitted on this channel B are transmitted (319) on this channel B line X25. Again, we see that an optimization is made, to limit the number of open communications simultaneously. If no X25 communication is open (320), one checks (321) whether a B-channel

  is established. In this case (322), the new stream is transmitted on this same channel (323).

  If not (324), a channel B is open (325), and this channel B is transmitted (326). In general, the router therefore seeks to implement a single communication line, and selects the one to be used. the most efficient for all flows to be transmitted. This principle is implemented continuously, and not only at the beginning of a stream. Thus, when a flow is completed, it is verified that, for the remaining flows, the current communication is the most adequate. Otherwise, there is an automatic switch to another line. Still with the aim of limiting the costs of communications, the router of the invention advantageously comprises data compression means, as well as

this is illustrated in Figure 2.

  If the stream is transmitted over a circuit-switched network 24, means 241

  conventional compressors compress the data before they are transmitted (242).

  Optionally, several compression methods can be implemented. We then choose the most efficient, after comparison, or an automatic selection is made in

depending on the type of flow.

  If the stream is transmitted on a packet-switched network 23, the comparison means 231 are advantageously of the type described in the French patent application FR 92 01657 of February 11, 1992 in the name of the same Applicant. According to this particularly efficient technique, the data fields of the packets to be transmitted are compressed, concatenated with data making it possible to locate the beginning of each packet.

  initially, and then cut into the format of the packets provided for the communication network.

  On receipt, the reverse operations are performed to restore the original packets. Thus, the number of packets actually transmitted is less than the number of

source packages.

  In addition, it should be noted that other criteria than the cost of a communication can be taken into account for the selection of one network or the other. In particular, the following may be taken into account, depending on the case: - the one-way or two-way aspect of the communication; the continuous or episodic aspect of communication; the response time of said communication networks; - the duration of the communication; - the speed of communication;

  - the communication resources available.

  Finally, the device of the invention is advantageously fully programmable.

  It is the user who associates to a type of flow and / or a destination a given network (or several, with an order of preference). Similarly, the user can force or prohibit

  certain situations, when necessary.

1l

Claims (10)

  Interconnection device (111, 112, 113) of at least two remote stations (141, 142) and / or local area networks (131 to 134) via at least two communication networks (121, 122), comprising on the one hand at least one first packet communication network (121) and on the other hand at least one second circuit-switched communication network (122), characterized in that it comprises means ( 25) for recognizing the type of an application flow to be transmitted, and means (22, 28) for automatically selecting one of said communication networks for said communication to be established, as a function of said type of flow. 2. Device according to claim 1, characterized in that said means (28) of selection take into account at least one of the criteria belonging to the group comprising: - the one-way or bidirectional aspect of the communication; - the continuous or episodic aspect of the communication; the response time of said communication networks; - the duration of the communication; - the speed of communication; - available communication resources; the billing costs and conditions associated with each of said communication networks.   3. Device according to any one of claims 1 and 2, characterized in that   uses an integrated services digital network (ISDN) comprising at least one digital information transfer channel (B) and at least one signaling and packet transfer channel (D), and in that said means (28) selection means select at least one of said channels according to said type of application flow, said channel D corresponding to said first   network and said B-channel corresponding to said second network.   4. Device according to any one of claims 1 to 3, characterized in that   said selection means (28) also take into account the communications (29) already established between the same stations and / or local networks, so as to use if possible the same communication resource.   5. Device according to any one of claims 1 to 4, characterized in that   includes means (231, 241) for compressing data associated with at least one said communication networks.   6. Device according to claim 5, characterized in that it comprises means   (241) compressing data transmitted on said first network.   A method for interconnecting at least two remote stations and / or LANs via at least two communication networks, comprising on the one hand at least one first packet communication network and on the other hand at least one second circuit-switched communication network, characterized in that it comprises the following steps: recognition (25) of the type of application flow to be exchanged; automatic selection (28, 22) of one of said communication networks   for said application flow, depending on the type of flow.   The method according to claim 7, characterized in that said automatic selection step comprises, for at least certain types of flows, a hierarchical selection procedure among at least two types of communication networks, as a function of networks already used.   9. Method according to any one of claims 7 and 8, characterized in that   includes a communications optimization step, aimed at circulating at least   two application flows of different types on the same communication line.   10. Process according to any one of claims 7 to 9, characterized in that   said types of application flows include: - file streams, corresponding to file transfers; - interactive flows, corresponding to interactive data exchanges; and   - the default flows corresponding to the other flows.   1. Process according to claims 9 and 10, characterized in that it comprises the   following steps: - if the stream is a file stream and no path to the recipient is open, opening (325) a path on a circuit-switched communication network; - if the stream is an interactive stream and no path to the recipient is open, opening (312) a path on a packet communication network; - if the stream is a file stream and a path to the recipient is open on a packet communication network, opening (318) a path on a circuit-switched communication network and closing (317) the path open on a packet communication network; - if the stream is an interactive stream and a path to the recipient is open on a packet switched communication network, maintaining (310) said path. The method as claimed in claim 11, characterized in that when a path is open for the transfer of at least two distinct application flows and an application flow.   is completed, the selection operation of an optimal communication network is repeated.   13. Process according to any one of claims 7 to 12, characterized in that   includes a programming step for associating at least one of said types of application flows and / or at least one recipient with at least one preferential communication network.