FR2980065A1 - Routing device for use in optical core of telecommunications network i.e. Internet, has sending unit configured such that flow of Ethernet frames emitted via channels corresponds to specific percent of maximum flow supported by channels - Google Patents

Abstract

The device (200) has a set of memory zones (MM1, MM2), and a set of ports (PS1, PS2) that is associated with the memory zones. Each port is connected to a transmission channel (C1, C2) of a telecommunication network i.e. Internet. One of the memory zones stores Ethernetframes intended for a set of routers (RO1, RO2). A sending unit (ME1) is configured such that flow of Ethernetframes emitted via transmission channels corresponds to 70 percent of maximum flow supported by the channels. Independent claims are also included for the following: (1) a data sending method (2) a computer program comprising instructions for performing the data sending method (3) a recording medium comprising a set of instructions for performing a data sending method.

Description

BACKGROUND OF THE INVENTION The invention relates to the general field of routers used in the core of optical telecommunications networks and more particularly relates to the detection of errors or faults in a WDM or optical fiber network. . Historically, SDH technology (for "synchronous digital hierarchy") was the type of multiplexing protocol typically used to support data traffic in the core of telecommunications networks such as the Internet for example. This WAN-type technology is here described with reference to an exemplary implementation represented in FIG. 1. More particularly, FIG. 1 represents an SDH-type network 1 comprising two high-speed routers 4 and 12 interconnected by one another. This TRANSI_ channel here comprises the following elements a first regenerator 6 connected to the router via an optical fiber F2, an ADM module 8 (for "add-drop n7") connected to the transmission channel (or conduit) TRANSI. regenerator 6 via an optical fiber F4, a second regenerator 10 connected to the ADM module 8 via an optical fiber F6, and an optical fiber F8 connecting the second regenerator 10 to the second router 12.

In operation, the router 4 is able to receive data in the form of Ethernet frames from the terminal 2, to unencapsulate (or extract) the IP packets therein and to re-encapsulate (or reinsert) the latter in new Ethernet frames. in order to route them to the second router 12 via the transmission channel TRANSI. The ADM module 8 supervises the transmission of data along the transmission channel TRANSI_ and in particular controls the different types of frames conveyed in the channel. The ADM module 8 is also able to maintain a connection table between the different terminals that can be located on either side of the transmission channel. The regenerators 6 and 10 are intended to amplify the optical signals transmitted via the transmission channel TRANS1 from the first router 4 to the second routobr 12.

The second router 12 is capable of routing the IP routes from the first router 2 to the remote terminal 14. The SDH technology is enabled to use the frames to detect failures. int don a punch. regenerators 6 and 10 are able to add information relating to a failure in a supervisory frame transmitted by another device of the SDH network 1. When a device of the SDH network receives a supervision frame, it decodes it and deduces 'a code contained in the frame the place where the failure occurred in the network.

However, Ethernet technology has massively spread in recent years in optical network cores because it allows very significant gains in terms of costs including (in the order of 60% on the price of interface cards routers) . This LAN technology has many advantages but is not compatible with the use of supervision frames. It is not possible to use the supervisory frame mechanism to locate or even detect a failure in a network using Ethernet. FIG. 2 represents the example of a type VVDM network ("wavelength-division / n") denoted 100 using Ethernet technology. More specifically, the WDM network 100 comprises two high-speed routers 104 and 112 connected to each other via a transmission channel (or channel) TRANS2. This channel TRANS2 here comprises the following elements: a first regenerator 106 connected to the router via an optical fiber F12, a VVDM module 108 connected to the regenerator 106 via an optical fiber F14, a second regenerator 110 connected to the WDM module 108 via an optical fiber F16 and a fiber optic F18 connecting the second regenerator 110 to the second router 112.

The router 104 is able to route, to the remote router 112, the IP data received in optical form from the client terminal 102. To do this, the remote router 112 de-encapsulates (or extracts) the IP packets in each Ethernet frame. received and, once the IP packets recovered, re-encapsulates (or insert) them into new Ethernet frames that the router 104 generates for this purpose. Each Ethernet frame thus generated comprises an IP packet transmitted by the client terminal Ti and the MAC address of the equipment to which this data is intended (i.e., the address MA [of the router RO1 for example). The regenerators 106 and 110 make it possible to amplify the optical signals transmitted by the transmission channel TRANS2 between the first router 104 and the second router 112. These readers, however, are optional: a direct connection of the routers 104 and 112 to the module 100. . posbible attenuatiol signal is limited, e along the transmission channel (case where distance course is short). Unlike the modul, -, ADM 8, the VVDM module 108, able to generate i ~ m / t therefore not u; .., other elements of the network VIDV 100 100 tr nt35 - one of the routers 104 and 112, and in particular its input / output card called communication interface card (the card then generates frame errors) and / or its optical module (the module then generates physical errors). The input / output card and the optical module will be described in more detail later in the context of the invention; one of the regenerators 106 and 110; and / or one of the optical fibers F12 to F18 collectively referred to as F (poor quality of the fiber, non-conforming radius of curvature, pinched fiber, etc.). In addition, undesirable signal losses can result in particular from: - the presence of dust or poor locking at optical connectors, - an inappropriate path length due to an erroneous theoretical calculation; and / or an optical signal that is too weak at the level of the optical module of a router. There is therefore today a need for endurance testing in a simple and almost instantaneous manner of optical telecommunication networks, of a similar type. VVDM in particular. In particular, there is a need to evaluate the quality of transmission lines connecting two high-speed routers in an optical telecommunications core network. OBJECT AND SUMMARY OF THE INVENTION To this end, the invention relates to a routing device in a core optical telecommunications network, providing the interface between at least one terminal and at least one router in the telecommunications network, the device comprising: at least one memory area and a dedicated port associated by configuration to that memory area, each port being connected to a respective transmission channel of the network; configuration means able to dedicate means for sending the device to said at least one memory zone for sending a stream of Ethernet frames included in said at least one memory zone via the transmission channel of the corresponding port; generating means adapted to generate first Ethernet frames for said at least one router; and receiving means adapted to generate second Ethernet frames from ul.-Jris Ethernet frames from said at least one terminal, and, where configuration is apt, for each .: Tn-er urn i-rp.:Triory to memorize p, ei1c net Elher frames destined for the rth (.1.1 '^ .. 1 1 oins -, memo: r - 4.1 L 1 UL, L! Z via at least one of said transmission channels corresponds to at least 70% of the maximum bit rate that can be supported by said transmission channel The optical telecommunications network corresponds, for example, to a VVDM type network or to an optical fiber network that is devoid of equipment of type VVDM.

The routing device according to the invention is advantageous in that it is capable of providing both conventional routing functions in an optical telecommunications network and, in a parameterizable manner, endurance testing of one or a plurality of transmission paths. transmission to which it is connected via its output ports. Thanks to the invention, it is possible to evaluate the transmission quality of transmission conduits in an optical telecommunications network and without modifying the architecture of current routers. Advantageously, the device may, if necessary, be remotely activated in order to perform one or more endurance tests of transmission paths of an optical telecommunications network. The invention eliminates the physical interventions that are usually required to test a conduit in transmission. The invention is also advantageous in that it makes it possible to carry out an endurance test of a transmission channel of an optical telecommunications network in order to deduce its quality (or reliability) in transmission and this, without soliciting processor-level resources of the routing device in particular. This is possible because OS1 level 2 Ethernet frames are generated and not OSI level 3 packets. The invention makes it possible not only to detect a fault or errors occurring on the transmission path between two routers of a core optical telecommunications network but also failures affecting the interface card of one of the routers (at the level of optical module of the interface card in particular).

According to a particular embodiment, the device comprises N first memory areas, N being a natural greater than or equal to two, in which the generation means are able to generate N sets of said first Ethernet frames intended for N respective routers out of said least one router, the device being configured to store each of the N sets of first frames in respectively one of said N first memory areas. mode of réalization advantageously makes it possible to simultaneously test | o quality in = r-lc, n-1: ssion of a pirii conduit transmissions connected to the device routing fim er ertion in the telecommumic.dtions optical network . The device can be so called: memory and transmission in the optical telecommunications network while routing client traffic from client terminals. In a particular embodiment, the configuration means are able to switch the device of a first mode where only the first frames are stored in the first memory areas, to a second mode where only the second received frames are stored in the second frames. memory areas, the configuration means being also able to switch the device of the second mode to the first mode. The device of the invention is thus capable of switching between a "test" mode allowing endurance qualification of transmission lines and a "router" mode ensuring the routing of client traffic in the optical telecommunications network. The inter-frame delay between each first Ethernet frame of the frame transmission sent by the sending means is preferably less than 10ps, and preferably equal to 0. A low or even zero inter-frame delay advantageously makes it possible to test a transmission duct at a flow rate close to the maximum flow rate that can be supported by the duct and thus to effectively evaluate the transmission quality of the duct. In addition, said first Ethernet frames preferably have a size of 4484 bytes. Setting the size of the first Ethernet frames as much as possible makes it possible to limit the resource demands on the processor of the device of the invention. Correlatively, the invention relates to a sending method implemented by a routing device of a core optical telecommunications network, the routing device performing the interface between at least one terminal and at least one router in the network, the device comprising at least one memory zone and a dedicated port associated by configuration to this memory zone, each port being connected to a respective transmission channel of the network, and receiving means able to receive second Ethernet frames 25 from said at least one terminal, the method comprising: - a configuration step for dedicating means for sending the device to said at least one memory zone; generating first Ethernet frames for said at least one router; and - generating second Ethernet frames from data included in Ethernet frames from said at least one terminal; storage of the first Ethernet frames intended for said routers in a first zone T 2 at each of the routers, and Ethernet trams that are inemore in at least one detixiome / orc rnemoiru different moiru- transmission channels corresponding to at least 70% of the maximum rate can be supported by the transmission channel. It will be noted that the different embodiments of the sending method according to the invention have the same advantages as those mentioned above with reference to the respective embodiments of the routing device according to the invention. The optical telecommunications network corresponds, for example, to a WDM type network or to an optical fiber network that does not have VVDM type equipment. According to a particular embodiment, the sending method further comprises a switching of the device of a first mode where only the first frames are stored in the first memory areas, to a second mode where only the second frames received are stored in the second memory areas. Alternatively, the switching step is from the second mode to the first mode. The inter-frame delay between each first Ethernet frame of said train is preferably less than 10ps, and even more preferably zero.

In addition, the first Ethernet frames preferably have a size of 4484 bytes. In a particular embodiment, the various steps of the sending method are determined by instructions of computer programs. Consequently, the invention also relates to a computer program on an information medium, this program being capable of being implemented in a router or more generally in a computer, this program comprising instructions adapted to the implementation of implement steps of a sending method as described above. This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other form desirable shape.

The invention also relates to a recording medium or information carrier readable by a computer, and comprising instructions of a computer program as mentioned above. The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise a storage means, such as a ROM, for example a CD ROM or a micro-circuit ROM: ect_ron.que, or a recording medium MagneU: lt .., for example a disqu, "py disc) or a hard drive. On the other hand, the USSR; information can e-e uo suppo-t. A further embodiment of the present invention will be described in the following description, which may be further described in the following description: below, with reference to the accompanying drawings which illustrate an embodiment having no limiting character. In the figures, FIG. 1 schematically represents an example of a known SDH network; FIG. 2 schematically represents an example of a known VVDM network; FIG. 3 schematically represents the architecture of a routing device according to a particular embodiment of the invention; FIGS. 4A, 4B and 4C show, schematically, three examples of operation of the routing device of FIG. 3; and FIG. 5 represents, in the form of a flowchart, the main steps of a sending method according to a particular embodiment of the invention.

DETAILED DESCRIPTION OF ONE EMBODIMENT The invention relates to a router implemented in the core of an optical telecommunications network, this router having the capacity to detect errors or faults occurring in the network. We now consider a router 200, according to a particular embodiment of the invention, implemented in a network VVDM 220 similar to the network WDM 100 of Figure 2. The router 200 and the method of sending it The following are described below with reference to a WDM type optical network. However, it will be understood that the invention can be applied to other types of optical networks, that is to say to optical telecommunications networks without VVDM type equipment. For example, the invention can be implemented in a network comprising simple optical fibers between each router. Figure 3 schematically shows the architecture of the router 200 of the invention. In the example envisaged here, the router 200 presents the hardware architecture of a conventional high-speed router such as the router 104. The router 200 according to [invention differs only the way in which it is configured. More specifically, the route here includes a processor 202, a ROM type ROM 204, a nonvolatile memory 27 and an interface card qt.ree, including instructions for executing the steps (denoted Ell at E13 and E21 to E24) of a method of sending the invention. The interface card 210 has Ethernet PE input ports and Ethernet PS output ports. The card 210 allows the reception of data from client terminals on a respective input port and the sending of optical Ethernet frames via transmission channels of the VVDM network 220 connected to a respective output port. As shown in FIG. 3, the interface card 210 makes it possible, for example, to route, via the transmission channel C1 connected to the output port PS1, the IP data originating from the client terminal T1 to a remote router R01.

The operation of the router 200 according to the invention is now described with reference to FIG. 5 and the three examples of operation represented in FIGS. 4A, 4B and 4C. In the example of FIG. 4A, the router 200 comprises two memory zones MM1 and MM2 corresponding to two different zones of the random access memory 206, these zones being used as buffers. It will be understood in the light of the following that the number of buffer zones is not limited to 2 and depends on the operating mode envisaged. Alternatively, at least one of the memory areas MM1 and MM2 may be located in a memory (not shown) of the communication interface card 210. Each of the memory areas MM1 and MM2 is associated with a respective output port PS1 and PS2 (and the corresponding sending means ME1 and ME2, see below). As indicated in more detail later, it is also necessary to reconfigure the mappings between the memory zones and the output ports so that different memory zones are assigned to the output ports of the router 200. In addition, each output port PS1 and PS2 is connected to a separate transmission channel (denoted respectively C1 and C2) in the VVDM network 100. The channels C1 and C2 can convey optically in the VVDM network 220 data, and in particular Ethernet frames. These transmission conduits C1 and C2 correspond for example to an optical fiber or to conduits such as TRANS2 described in FIG. 2. The interface card 210 of the router 200 also comprises sending means denoted ME1 and ME2 respectively associated with the zones memories MM1 and MM2 and connected to the respective output ports PS1 and PS2. The interface card 210 furthermore comprises receiving means RE1 capable of breaking: "the IP data contained in each Ethei frame i-1 (, t received (ii-1 r) from the terminal 11. nyyyt P71 in the title For example, the configuration means [ONE to transfer, in the buffer zone (ie MM1 in this example) associated with the port PS1, the client Ethernet frames generated. The RE1 receivers thus take as input parameters the Ethernet frames from the client terminal T1 and an output port whose value (or an identifier) is provided by the configuration means CONF. The interface card 210 further comprises means MG generation configured to generate Ethernet frames based on: the AD2-rated MAC address of the RO2 router to which the generated Ethernet frames are to be sent, and the output port to be used (ie PS2 in this example) For this purpose, the generation means therefore receive as input the address AD2 and the value (or an identifier) of the port PS2 from the configuration means CONF. The generation of the Ethernet frames by the MG means is preferably pseudo-random from a polynomial calculation. Each Ethernet frame generated by the means MG notably comprises the MAC address of the destination equipment (i.e., the address AD2 of the router RO2 in the present case). As described in more detail later, the Ethernet frames generated by the MG generation means make it possible to carry out a test as to the quality of transmission offered by the transmission channel C2 of the WDM network 220. This type of Ethernet frames will therefore be thereafter called "Ethernet test frames". The generation means MG are configured to transmit the test Ethernet frames thus generated to a memory area (MM2 in this example) which has been specifically assigned to them by the configuration means CONF. Typically, the configuration means [ONE correspond to a human-machine interface comprising a keyboard for example and allowing the operator to specify the output ports that he wishes to test endurance or, alternatively, that he wishes to use to transmit customer data in the VVDM network. An exemplary implementation of the sending method by the router 200 is now described with reference to the example of FIG. 4.

During operation, the router 200 can receive on the port of er.Lo PF1 c'it 'of the terminal üent Tl (step Ell). This data is received in the form corresponding to Ethernet frames in which are encapsulated. The MAC address AD1 is for example provided by the configuration means CONF to the reception means RE1. On order of configuration means CONF, the receiving means RE transmit the client Ethernet frames thus generated in the memory area dedicated to the output port PS1, namely the memory area MM1. The reception means RE1 thus take as input parameters the Ethernet frames from the client terminal T1 and the output port by which the new client frames must be sent. The memory zone MM1 stores (step E12 of "buffering"), as it receives them, client Ethernet frames (denoted TR11, TR12 ...) transmitted by the reception means RE1. The associated sending means ME1 send sequentially, on the transmission channel C1 of the output port PS1, client Ethernet frames stored in the memory area MM1 (step E13). A client Ethernet frame stream is thus transmitted, via the transmission channel C1, to the destination router R01. A frame stream corresponds to a succession of at least two frames having a determined inter-frame delay. In parallel, the router 200 is configured via the configuration means CONF to test the transmission quality of the transmission channel C2 connected to the output port PS2. For this purpose, the configuration means CONF assign (assignment step E21) a zone MM2 of the volatile memory 206 to the output port P52 to which the router RO2 is connected via the transmission channel C2. In addition, the CONF configuration means provide the MAC address AD2 and an identifier (or the value) of the output port PS2 to the generation means MG. Upon receipt of these input parameters (and possibly upon receipt of a predetermined trigger command initiated either manually or automatically), the generation means MG generate (steps E22) Ethernet test frames at a rate predefined by configuration. Each of these test frames contains in particular the MAC memory AD2 of the router RO2 to which the frames are intended. The associated buffer memory MM2 stores (step E23 of "buffering") the ethernet frames of tc.st as they are sent and sent by the generation means MG. The means d. associated send E24) then successiveme.nt, on the transmiççien channel (7.7 noit output PS2, truntes test Ethernet (denoted TR21, TR22 ...) or iryn MM2 .There a test Ethernet transmission nt thus transmitted, à k 120 120 120 120 120 120 du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du du The sending of such a rate on the conduit C2 can test it in endurance (or "load"). Only a large flow of frames along the duct makes it possible to evaluate the transmission quality offered by the duct C2, in order to guarantee a flow rate of 70 "ib or more of the maximum flow rate of the duct C2, the means of generation MG and the means ME2 send are shown to emit at the output of the router 200 a test Ethernet frame train respecting a delay inter-frame LED (ie the time interval between each test transmission Ethernet frame) and a frame size Appropriate TAI By playing on the LED and TAI parameters, it is possible to guarantee a certain "PPS" rate of IP packets per second.Using the charts provided by the router manufacturers, it is then possible to ensure that the PPS provided the router corresponds to a FPS rate of Ethernet frames per second greater than or equal to 70% of the maximum rate that the transmission path C2 can provide, In this example, the generation means and the sending means of the invention. his t configured so that the size of the test Ethernet frames generated and sent on the transmission path to be tested is large, and preferably equal to the maximum size allowed by the Ethernet protocol, namely 4484 bytes. Furthermore, the chosen LED inter-frame delay is preferably less than 10ps, and is ips for example. Even more preferably, the inter-frame delay LED is zero so that the Ethernet test frames are sent without waiting one after the other on the transmission channel tested. Moreover, the router RO2 processes each test Ethernet frame received (because these frames contain its AD2 MAC address) and determines in particular the value contained in the CRC field of each frame in order to check whether an error occurred during the transmission along of the transmission duct C2. The determination of the CRC parameter allows the router R02, for each frame, to determine if an error has been introduced into the data of the frame. For each detected error, the router RO2 increments an error counter, the level of this counter being representative of the transmission quality of the transmission path C2. The router 200 is thus capable of routing client data to a first output port PS1 and, simultaneously or sequentially, endurance testing the transmission path connected to another output port PS2. In Figure 4B, Id --outeur 200 r! Lddre in rd nue the possible rvennne nmven: ln 1. "15m, ses-eceptIon RE I rnt: 1110Eru (cipres h th.:1 indyens RE1 this ayens -nfia itior rot.i- test in endurance the transmission path Cl connecting the output port PS1 to the remote router RO1 and this, while routing the possible client Ethernet frames to the router R02. By making this type of configuration changes, it is possible to advantageously endurance test sequentially the different transmission paths while ensuring a routing of any customer data from the terminal T1. In a particular embodiment, the router 200 is able to switch, on command of a user or in accordance with at least one predefined criterion, from a first operating mode to a second operating mode, and vice versa, according to the first mode, only test Ethernet frames are stored in memory zones dedicated to each rt output from the router. In the second mode, only the client Ethernet frames received by the router are stored in the dedicated memory zones. Note also that the router according to the invention may also be able to endurance test a plurality of transmission conduits simultaneously. The configuration means CONF then assign a memory area separate from the memory 206 for each output port on which it is necessary to endurance test the corresponding transmission channel. In other words, a memory zone is dedicated to each of the remote routers (R01, R02, etc.) placed at the end of the transmission conduit for endurance testing of said conduits. More generally, the CONF configuration means make it possible to define, for each output port, the operating mode that the router 200 must adopt, namely a client data transmission mode (or "client traffic") or a endurance test mode of the corresponding transmission conduit. FIG. 4C represents the case where two transmission conduits are endurance tested simultaneously. More particularly, the router 200 differs here in that it is configured to test in endurance both the transmission channels C1 and C2.

In this example, the interface card 210 comprises generation means MG1 and MG2 capable of generating test Ethernet frames. The generation means MG1 (respectively MG2) receive as input the MAC address AD1 (respectively AD2) of the remote router RO1 (respectively R02) intended to receive the test frames. In the same way as the generation means MG represented in FIGS. 4A and 4B, the generating means 1 \ 1G1 and MG2 transmit the Ethernet terminals generated in a dedicated memory area (MM1 respectively to ports PS1 and PS2) (and to the remote routers and transmission channels The mempw-tion Linn: the memory areas i'its11 MM2, the sending MEI frames ME7 c <ie test of dualif: caL: one cndmancæau router in the network VVDM 220. If this is In this case, the IP data originating from the client terminal are encapsulated in client Ethernet frames, these frames then being stored in dedicated memory areas of the memory 206. These memory zones are therefore distinct from the memory zones assigned to the storage of Ethernet frames of the memory. test.

However, the router 200 does not necessarily receive client Ethernet frames from the terminal T1 when it tests in endurance at least one of the transmission paths of the VVDM network 220. In a particular embodiment, the configuration means CONF configure the minus one generation module MG and one reception mode RE so that they respectively transfer test Ethernet frames and client Ethernet frames in the same memory area MM of the random access memory 206. This particular mode makes it possible at the same time to endurance testing a transmission path and router, via the same conduit, IP data from a client terminal. In general, the router according to the invention is advantageous in that it is capable of providing both conventional routing functions in an optical telecommunications network (of the VVDM type for example) and, in a customizable way, to test in endurance one or a plurality of transmission conduits to which it is connected via its output ports. It is preferably determined in advance whether an output port is available (ie not assigned to the routing of client traffic) in order to know if the corresponding transmission path can be qualified in endurance. As indicated above, the invention does not exclusively concern WDM networks but can be applied to any optical network. The router of the invention may in particular be implemented in an optical core network comprising at least one optical fiber.

Thanks to the invention, it is possible to evaluate the transmission quality of transmission paths in a WDM network without changing the architecture of current routers. Once installed in a conventional router, the PROG program makes it possible to execute the sending method according to the invention. Advantageously, the method of the invention may, if necessary, be triggered remotely by the operator of the optical telecomunications network, of the VVDM type, for example. The invention makes it possible to dispense with the physical interventions that are usually necessary to test a conductor in a trnnsmi7sfon. Invenuan rent dvantac, Ieus, .2 cry what the peripct; The invention makes it possible not only to detect a fault or errors occurring on the transmission path between two routers of a core of the invention. optical telecommunications network but also failures affecting the interface card of one of the routers (at the level of the optical module of the interface card in particular) .The present invention finds more particularly an application in the implementation of test transmission endurance in high-throughput routers operating in \ A / DM or fiber-optic network cores A "high-speed router" is understood to mean a router capable of processing at least 1600 GB bits / s, c ie 16 links at 100 GB bits / s or 160 links at 10 GB bits / s for example.

Claims (2)

REVENDICATIONS1. Routing device (200) in a core network of CLAIMS1. A routing device (200) in an optical telecommunications core network (220), providing the interface between at least one terminal (Ti) and at least one router (R01, R02) in said network, the device comprising - at least a memory zone (MM1, MM2) and a dedicated port (PS1, P52) associated by configuration with this memory zone, each port being connected to a respective transmission channel (C1, C2) of said network; configuration means capable of dedicating sending means (ME1, ME2) from said device to said at least one memory zone for sending a stream of Ethernet frames included in said at least one memory zone via the transmission channel of the corresponding port ; generation means (MG1, MG2) capable of generating first Ethernet frames for said at least one router; and receiving means (RE1) adapted to generate second Ethernet frames from data included in Ethernet frames from said at least one terminal, and said configuration means (CONF) being able to assign, for each of said routers, a first memory area (MM2) for storing the first Ethernet frames for said router, the second Ethernet frames being optionally stored in at least a second memory area (MM1) different from said first memory areas, the generating means and the means for sending associated with each of said first memory areas being configured so that the rate of the stream of Ethernet first frames transmitted via at least one of said transmission channels corresponds to at least 70% of the maximum rate that can be supported by said transmission channel. 2. Device according to claim 1 comprising N first memory areas, N being a greater or equal natural in which the generation means are able to generate N sets of t: rare Ethernet for N respective routers of said mo: r15 A router , di-rn -, 'tif being cor, icHr-, to memorize each of N sets of first frames res [. At least one of said second received Ethernet frames is stored in said second memory area. 4. Device according to any one of claims 1 to 3, wherein the configuration means are adapted to switch the device of a first mode where only the first frames are stored in said first memory areas, to a second mode where only the second received frames are stored in said second memory areas, the configuration means being also able to switch the device of said second mode to said first mode. 5. Device according to any one of claims 1 to 4, wherein the inter-frame delay between each first Ethernet frame of said train is less than 10ps. 6. Device according to claim 5 wherein said inter-frame delay is zero. 7. Device according to any one of claims 1 to 6, wherein said first Ethernet frames have a size of 4484 bytes. 8. Method of sending implemented by a routing device (200) of an optical telecommunications core network (220), said routing device providing the interface between at least one terminal (11) and at least one a router (R01, R02) in said network, said device comprising at least one memory zone (MM1, MM2) and a dedicated port (PS1, PS2) associated by configuration with this memory zone, each port being connected to a transmission channel (Cl, C2) respective of said network, and receiving means (RE1) adapted to receive second Ethernet frames from said at least one terminal, the method comprising: - a configuration step (E21) for dedicating means for sending (ME1, ME2) said device (200) to said at least one memory area; the generation (E22) of first Ethernet frames intended for the said at least one router; and - generating (RX) second Ethernet frames from data included in tramlines, nhernel from said at least one terminal; - The redirection (E23) of the first Ethernet frames to said routers in a first .., o! - memory (f \ IN12) assigned nor each of said rouLeurs. In the case of Ethernet frames, the following factors shall be used: 27, (E.3) at least one second zone, one of said transmission channels shall correspond to at least 70% of the maximum rate that can be supported by said transmission channel 9. The method of claim 8 further comprising switching the device from a first mode where only the first frames are stored in said first memory areas to a second mode where only the second received frames are stored. in said second memory areas, or switching from said second mode to said first mode 10. The method of claim 8 or 9, wherein the inter-frame delay between each first Ethernet frame of said train is less than 10ps. The method of claim 10, wherein said inter-frame delay is zero 12. The method of any one of claims 8 to 11, wherein said first Ethernet frames. The computer program (PROG) comprising instructions for executing the steps of the sending method according to any one of claims 8 to 12 when said program is executed by a computer. A computer-readable recording medium (204) on which a computer program (PROG) is recorded including instructions for executing the steps of the sending method according to any one of claims 8 to 12. 25