Switching method, terminal and corresponding computer program

1. Field of the invention

0001The field of the invention is that of radio frequency communications, and more particularly of the management of frequency resources between terminals or stations, and an access point to a communication network. The invention applies in particular, but not exclusively, to equipment of a wireless local area network, and for example to systems emitting in a frequency band located around 2.4 GHz or 5 GHz, in which, in particular, equipment according to the IEEE 802.11n standard or its revised versions, also called Wi-Fi.

0002Here we mean by equipment including an element belonging to a set of basic services (in English BSS for <i>"Basic service set"</i>), formed by an access point (in English AP <i>"Access point"</i>), or a wireless personal network coordinator (PNC for <i>"PicoNet Coordinator"</i>), and the stations associated with this access point, that is to say the stations located in the coverage area of ​​this access point.

0003More specifically, the invention relates to the management of resources of an access point capable of transmitting simultaneously on at least two channels, and more particularly a tilt mechanism of a terminal between two channels.

2. Prior Art

0004The following describes the management of the frequency resources of equipment of a network operating notably according to a Wi-Fi technology as defined in the IEEE 802.11r standard proposing a fast transfer (in English). <i>"Fast BSS transition"</i>) a station between a first set of basic services managed by a first access point and a second set of basic services managed by a second separate access point.

0005According to this standard, an access point manages a single given channel. A mechanism is proposed to allow a switchover of a data transmission from one access point to another access point, and therefore in general with a change of channel in the event of mobility of the terminal. In other words, this IEEE 802.11r standard makes it possible to switch over a data transmission, the terminal concerned then having to associate with a new access point managing a new channel.

0006This failover mechanism comprises several processing phases, which are long and complex. Indeed, the tilting terminal must "discover" the new access point, to authenticate and associate with this access point including exchange and recognition of encryption keys.

0007Despite the gains made by this technique where the duration of preparation of the transition (up to a few seconds) is anticipated, there is nevertheless a transition time of the order of 50 ms. Such a switching time introduces a momentary interruption of service, which is not acceptable for many applications, and for example for the transmission of a video.

0008In a case of disturbance of the radio channel by a radar, it has been envisaged that an access point can change channels in the same frequency band, to avoid disturbing the radar transmissions considered as priority. Indeed, in this case, a channel switchover can be implemented by the same access point. For this purpose, the 802.11h standard specifies, in particular, adaptations for the management, in Europe, of spectrum and transmission power relating to computer networks using a wireless link in the 5 GHz frequency band. The standard provides, for example, mechanisms for dynamic frequency selection and control of the transmission power.<i>Spectrum and Transmit Power Management Extensions in the 5 GHz Band</i> ").

0009In particular, according to this standard, the dynamic frequency selection mechanism (DFS for <i>Dynamic Frequency Selection</i> ") Is based on a failover management frame that announces the channel change (in English" <i>Channel Switch Announcement</i> "). Thus, an access point deciding to launch a channel switching procedure, sends this specific management frame with access to the priority channel (free channel for a specific duration noted "PIFS" for "<i>Point coordination function Inter-Frame Spacing</i> " in English). This frame contains in particular the number of the new channel and the time to count down before the change. This time is expressed in number of "beacons" (in English "<i>beacon</i> ").

0010A beacon is a management frame containing all the information of the communication network, in particular the information of the channel on which they are transmitted periodically by the access point. They play a role of stamping, and synchronize all the terminals of the same channel, attached to the access point. Time intervals (in English "<i>Beacon Interval</i> "), Which separate the transmission of two successive tags are set by the access point. In addition, these tags contain information that makes it possible to know the characteristics of the basic service sets proposed by the access point, for example the identity of the access point (BSSID).<i>Basic Service Set Identification</i> "), The frequency band, the channel number in this frequency band, the options supported by the PHY / MAC layers.

0011This standard offers the advantage of interrupting transmissions for a very short time, since the stations find the same access point in the new channel, without new authentication or new association. However, this approach only allows a channel change in the same frequency band.

0012In addition, this approach is not multichannel, because an access point is always, at a given time, associated with a single channel. Indeed, the disadvantage of this technique according to the 802.11h standard is that it causes channel switching of all terminals attached to the access point. In other words, if a channel change action is initiated, all the streams managed by the access point are diverted to another channel. It is not possible to distribute the load between different channels for a given access point.

0013The US patent application <patcit id="pcit0001" dnum="US2004185887A"><text>US 2004/185887</text></patcit> relates to a wireless network node in a multichannel context, the network comprising at least two transceivers set statically on non-interfering frequency channels. The disadvantage of this technique lies in the fact that at each channel change an association and authentication procedure must be performed, thereby lengthening the transmission interruption time during channel or band switching.

0014There is therefore a need for a solution to accelerate these channel changes for multi-channel access points for current projects. Indeed, a situation of asynchronous channels managed by different mechanisms is currently envisaged in the future IEEE 802.11ac standard currently under development.

0015This future standard aims at the use of radiofrequency channels of increasing width, having a width for example equal to 80 MHz (corresponding to the aggregation of four channels of 20 MHz) or more.

0016Thus, FIG. 1a illustrates this mechanism for a desired channel of 20 MHz, 40 MHz, 60 MHz or 80 MHz width, which can be alternatively constituted by four adjacent 20 MHz channels. To do this, this mechanism defines a primary channel on which a CSMA-CA access mode ("<i>Carrier Sense Multiple Access-Collision Avoidance</i> As described in 802.11-2007, paragraph 9.1 <i>"MAC architecture"</i>, 9.1.1 <i>"DCF"</i> is applied as well as the sending of a beacon by an access point which also manages a so-called secondary channel (<i>"Secondary Channel"</i> in English), a tertiary channel (<i>"Tertiary Channel"</i> in English) and a so-called quaternary canal (<i>Quaternary Channel</i> in English). According to this alternative, the CSMA-CA access mode is implemented only on the primary channel the other three channels being synchronized on this primary channel, that is to say operating in the same access mode.

0017However, this future standard provides that it may be difficult to find a free 80 MHz band in the available band and offers an alternative, illustrated by the<figref idref="f0001">figure 1b</figref>, to transmit at 80 MHz by creating two desynchronized and frequency-separated channels, which summed together forms an 80 MHz channel. So, according to this alternative represented on the<figref idref="f0001">figure 1b</figref>, the two desynchronized channels are 40 MHz channels each composed of a primary channel (1 and 2) and a secondary channel (1 and 2). On each of the primary channels 1 and 2, two separate access modes and separate tags are then used, the set being managed by a single access point which then has multichannel functionality.

0018In addition, the future standard IEEE 802.11ad, which is also in the process of standardization, concerns a Wi-Fi system operating at 60 GHz and introduces the possibility of switching between the 60 GHz frequency band and the 5 GHz frequency band. This future standardization also requires a fast and smooth switching without alteration of any transmission in progress, whether the channels at 60 GHz and 5 GHz are synchronized or not.

0019In this particular context concerning future standardizations, the inventors have therefore identified a need for a new technique that makes it possible to optimize the allocation of resources and the switching from one channel to another in a fluid and rapid manner, that the channels are synchronous. (ie, managed by the same access mode) or asynchronous (that is, managed by separate access modes from one channel to another), and that these channels are either non-contiguous.

3. Presentation of the invention

0020The invention proposes a new solution that does not have all of these disadvantages of the prior art, in the form of a data transmission failover method in a communication network comprising at least one access point exchanging data with at least one terminal attached to it and managing a set of at least two communication channels.

3.1 Resource allocation method (unclaimed)

0021According to a first approach (not claimed), the access point implements an association and authentication phase of a terminal with said access point, comprising a step of associating with said terminal a unique identifier. in said access point.

0022The access point also implements a management phase of said set of channels, using a reference table associating with a terminal, identified by its unique identifier, one of said channels. Said management phase further comprises the following steps:<ul><li>detecting a switchover of a terminal attached to said access point, from a first channel, said original channel, to a second channel, said destination channel;</li><li>updating said reference table to associate the unique identifier of said terminal with said destination channel.</li></ul>

0023Thus, in a multichannel access point, at least some parameters, and in particular the identifier of a terminal, are shared between all the channels associated with the access point. In other words, the identifier is independent of the channel used at a given moment. The association and authentication procedure of a terminal is not done, as is the case according to the prior art, with a particular channel, but with the access point (and therefore for all the channels that he manages). This approach makes it possible to speed up a channel change, since it is no longer necessary to launch a new association and authentication procedure with each change of channel. Thus, a channel change can be done in a sufficiently short time,

0024To manage the various communication channels that it supports, the access point has a reference table associating the unique identifier of a terminal with the current channel (or original channel) associated with this terminal. . This table, which should be interpreted here in the broad sense of a tool associating information (in the form of list, of database, ...), makes it possible to manage a change of channel for a particular terminal, without the other terminals associated with the access point are affected.

0025When the access point detects an effective failover (which can be performed following a request issued by the access point or following a decision of the terminal, according to the different approaches, as explained below) of a terminal on another channel, it updates this reference table accordingly.

0026According to a particular embodiment, said association and authentication phase also comprises a step of exchanging at least one encryption key between said access point and said terminal, where said encryption key (s) are common to each other. said set of channels.

0027Thus, according to this embodiment, the exchange of encryption keys between the access point and the terminal is valid regardless of the channel associated with the terminal. Thus, the encryption keys originally exchanged remain in effect after a switchover. The terminal being identified and authenticated by the access point and all the channels managed by it, it is not necessary to perform a new exchange of keys in case of failover, which also improves the tipping time.

0028According to one particular characteristic, the method further comprises, in said access point, a step of transmitting on said original channel a channel change command allocated to a terminal attached to said access point, said command allocated channel change system comprising at least said terminal identifier and an identifier of said destination channel.

0029Thus, according to this embodiment, it is the access point that decides the opportunity of a channel change for one of the terminals for which it is responsible, based for example on a load analysis. respective channels and / or the needs of the terminals. It then issues a specific allocated channel change command to the terminal, the assignments not being modified for the other terminals.

0030In other words, an access point may require a terminal to change channels, without necessarily the other terminals of the system transmitting on the same channel are forced to switch also.

0031In some implementations, it is of course possible that such a command simultaneously affects several terminals, or even if necessary all the terminals of a given channel.

0032For example, an access point detecting an overload on a particular channel may decide to change the channel only part of the terminals transmitting on that channel, so as to discharge the channel in question.

0033As already stated, this channel change can be done very quickly, since it does not require renegotiation of parameters at the destination channel, these being shared at the access point, that is to say common to all channels managed by this access point.

0034According to a particular aspect, the channel change control comprises a designation field of the terminals concerned, comprising a plurality of bits each associated with one of the terminals, and taking a first value when the terminal must perform the switchover to the channel. destination and a second value when it needs to keep the original channel.

0035Thus, a "bitmap" field for designation of the terminals concerned is added to the channel switching command, which enables quick and easy detection in the terminals.

0036According to another variant, the method further comprises a phase of synchronization of the origin and destination channels, comprising a step of simultaneous transmission of a first beacon on said original channel and a second beacon on said channel of destination.

0037By tag (in English " <i>beacon »</i>) reference is made in particular to a management frame (for example according to the IEEE 802.11 standard) which contains information of the communication network. For example, the information of a channel is the identity of the access point (BSSID).<i>"Basic Service Set Identification"</i>), the frequency band, the channel number in this frequency band, the options supported by the PHY / MAC layers. These tags are transmitted periodically by the access point on each channel. However, these channels are not always synchronized, the tags are not synchronized either.

0038According to this embodiment, therefore, synchronization of the beacons is provided, at least at the moment when a switchover is desired. Thus, the access point synchronizes at least the two channels concerned. In other words, the beacons of each channel are sent on each channel with the same start time.

0039In this way, when an access point requests a terminal to change channels, it does so just before the beacons are sent, so that the changing channel terminal receives the beacon of the original channel and the beacon destination channel immediately after it is switched to the destination channel. Thus, the channel change is accelerated because the changing channel terminal will not have to wait for the beacon of the destination channel to continue transmitting.

0040According to another embodiment, the method further comprises a step of transmitting, from said access point to at least one terminal connected to said access point, on said origin channel, at least one information relating to the state of at least one separate channel of said original channel.

0041In this embodiment, a terminal can decide for itself to change channels, without having received the order of the access point. Indeed, it is informed of the state, and in particular of the load and / or availability, of at least one other channel (preferably of all the channels associated with the access point), and can therefore determine if a tipping is necessary, or desirable, according to its current and / or future needs.

0042Advantageously, said at least one information relating to the state of at least one channel separate from said original channel is transmitted by encapsulation in a beacon of said original channel.

0043Thus, according to this embodiment, quasi-instantaneous channel switching is allowed. Indeed, on the original transmission channel, information is transmitted in the beacon of the original channel relating to the other channels, for example beacons (or part of the beacon information) for each potential destination channel. . Thanks to this encapsulation, each terminal is notified of the state of the original transmission channel but also of the state of the other potential destination channels. Accordingly, since the terminal has been previously notified of the state of the destination channel, the channel switching phase can be almost instantaneous.

0044In other words, the terminal is constantly aware of the characteristics of the potential destination channels. Thus, when it receives from the access point on the original channel an allocated channel change command, the terminal does not need to wait for the destination channel beacon previously synchronized with the channel tag. origin to switch to the destination channel. Indeed, unlike the prior art, it can switch immediately after receiving the allocated channel change command, since it simultaneously has the tag of the original channel and the destination tag.

0045According to this approach, it is also possible for the access point to ask a terminal at any time to change channels (by specifying the destination channel or by letting the terminal determine its destination channel), without necessarily waiting to transmit the channel change command allocated just before the beacons are issued. As a result, the waiting time of the destination channel beacon (synchronized with the beacon of the original channel) is saved.

3.2 Tipping method according to the invention

0046The invention relates to a method of switching data transmission from an original channel to a destination channel, in a terminal associated with an access point of a communication network, said access point managing a set at least two communication channels.

0047According to the invention, such a method comprises, in said terminal, an association and authentication phase with said access point, comprising a step of obtaining a unique identifier in said access point, and a failover management phase implementing the following steps:<ul><li>determining a destination channel managed by said access point, by processing at least one information received from said access point;</li><li>switching the data transmission from said original channel to said determined destination channel, without modifying said identifier.</li></ul>

0048Thus, according to the invention, a terminal can switch a current transmission on a destination channel, without renegotiation at the destination channel of a new identifier, and if necessary other parameters such as encryption keys.

0049From the processing of the information received from the access point, the terminal can directly perform the switchover, taking into account only the necessary synchronization information (tags).

0050According to one embodiment, the terminal detects the need to change channels (overloaded original channel or need more resources) and therefore decides to switch. It processes information received from the access point that allows it to know the channels managed by the access point also managing the original channel, which allows it to select an available destination channel. For example, the information received from the access point corresponds to the tags of each potential destination channel. The terminal then switches its current transmission to this selected destination channel.

0051According to a particular aspect of the invention, said information received from said access point comprises a channel change command allocated to said terminal, comprising at least one identification information of said terminal and an identifier of said destination channel.

0052Thus, the terminal thus receives a "switching" order, or allocated channel change command issued by the access point to its attention, and processes this command by first checking whether it is concerned by the command ( analyzing a "bitmap" field provided for this purpose, in a particular embodiment), and switches accordingly to the destination channel selected by the access point and specified in the switch command.

0053According to another variant, the method further comprises, in said terminal, the following steps:<ul><li>receiving, on said source channel, at least one information relating to the state of at least one other channel managed by said access point;</li><li>switching decision of said original channel to a destination channel, according to said at least one information relating to the state of at least one other channel.</li></ul>

0054This information can in particular be a beacon, or part of a beacon, of the other channels. According to a particular embodiment, it is transmitted encapsulated in a tag of the original channel, which simplifies the synchronization of the switchover.

0055With the aid of this information, the terminal can decide for itself, without having received an order from the access point, to perform a switchover, if it appears desirable or necessary, depending on its current and / or future needs.

0056In another embodiment, the invention relates to an access point in a communication network, exchanging data with at least one terminal attached thereto, and managing a set of at least two communication channels. Such an access point comprises means for associating and authenticating a terminal with said access point, associating with said terminal a unique identifier in said access point, and means for managing said set of channels, implementing a reference table associating with a terminal, identified by its unique identifier, one of said channels, and also comprises:<ul><li>means for detecting a switchover of a terminal attached to said access point, from a first channel, called the original channel, to a second channel, called the destination channel;</li><li>means for updating said reference table for associating the unique identifier of said terminal with said destination channel.</li></ul>

0057Another aspect of the invention relates to a terminal of a communication network, exchanging data with an access point to which it is attached, on a communication channel, called the original channel, selected from a set of least two communication channels, managed by said access point.

0058Such a terminal comprises means of association and authentication with said access point, comprising means for obtaining a unique identifier in said access point, and failover management means implementing:<ul><li>means for determining a destination channel managed by said access point, by processing at least one information received from said access point;</li><li>means for switching the transmission of data from said original channel to said determined destination channel, without modifying said identifier.</li></ul>

0059The invention also relates to a computer program comprising instructions for implementing a data transmission resource allocation method or a data transmission switching method as described above when this program is executed by a processor.

0060In addition, the invention relates to a signal exchanged between an access point and at least one terminal for implementing a data transmission resource allocation method or a data transmission failover method. as previously described. According to the invention, such a signal carries data selectively associating at least one terminal with at least one destination channel.

0061As specified above, the invention is characterized by the fact that the channel change is not necessarily made for all the terminals associated with a channel, but can be done specifically for one (or more) identified terminals.

0062Advantageously, such a signal further comprises at least one of the elements belonging to the group comprising:<ul><li>a field designating at least one terminal concerned by a channel switching command;</li><li>a beacon of an original channel encapsulating at least a portion of at least one beacon of at least one other channel separate from said original channel.</li></ul>

4.

List of Figures

0063Other features and advantages of the invention will appear more clearly on reading the following description of a particular embodiment, given as a simple illustrative and nonlimiting example, and the appended drawings, among which:<ul><li>FIGS. 1a and 1b, already commented on in relation to the prior art, illustrate two examples of known channel assignment mechanisms;</li><li>the <figref idref="f0002">figure 2</figref> presents the main steps of the allocation method implemented in an access point;</li><li>the <figref idref="f0002">figure 3</figref> discloses the structure of an allocated channel change control according to an embodiment of the invention;</li><li>the <figref idref="f0003">figure 4</figref> illustrates in detail the particular step of "synchronization" of two asynchronous channels of the allocation process represented in relation to the <figref idref="f0002">figure 2</figref> ;</li><li>the <figref idref="f0003">Figures 5a to 5c</figref> illustrate two variants of the invention for providing information relating to the state of at least one channel separate from the original channel;</li><li>the <figref idref="f0004">figure 6</figref> presents the main steps of the switching method in a terminal associated with an access point according to one embodiment of the invention;</li><li>the <figref idref="f0004">Figures 7a and 7b</figref> respectively illustrate the structure of an access point and a terminal according to one embodiment of the invention.</li></ul>

5. Description of an embodiment of the invention

5.1 General principle

0064The general principle of the invention is therefore based on the sharing of certain parameters, including the identifier of a terminal for all the channels associated with the same access point.

0065Thus, the association and authentication procedure of a terminal is unique with the access point and does not need to be repeated at each channel change.

0066As a result, a channel change can be made in a sufficiently short time without interrupting a current transmission on a changing channel terminal. In addition, such a channel change does not necessarily affect all the terminals attached to the access point transmitting on the same channel.

5.2 Case of the allocation of data transmission resources (unclaimed)

0067So as illustrated in <figref idref="f0002">figure 2</figref>, the method of allocating data transmission resources according to this embodiment is implemented in the access point AP managing for example 3 channels (C1, C2, C3). A first association and authentication phase 21 includes an association step 22 which consists in assigning to a terminal T a unique identifier AID. This identifier AID is shared by, or available to all the channels managed by the access point namely C1, C2, C3. Thus, the access point AP mutualizes the information relating to the identity of the terminal T for all the channels C1, C2, C3 that it has in management. In other words, even if the terminal T is placed on the channel C1, the channels C2 and C3 know the identity of this terminal T.

0068The access point AP can also manage one or more other terminals (not shown), independently optimizing the switching of each of the terminals on a given channel.

0069The association and authentication phase 21 may also comprise a step 23 for exchanging at least one encryption key K between the access point AP and the terminal T. Once this exchange has been made, the terminal T is identified and associated with the access point AP and the exchange of encryption key is valid regardless of the channel associated with the terminal T. In other words, the encryption key K exchanged is retained after a switchover of the terminal on a another channel managed by the AP access point. In the event of a switchover of the terminal T, it is therefore no longer necessary to carry out a new key exchange.

0070Thus, the single association and authentication phase 21 between the terminal T and the access point AP managing several channels makes it possible to optimize the switching time, since no reassociation or reauthentication of the terminal is necessary during the switching the terminal to a new channel managed by the AP access point.

0071The access point also implements a phase of management of all the channels by the access point AP. This management phase 24 includes a step of detecting a switchover of the terminal T attached to the access point AP from a first channel, called the original channel, to a second channel called the destination channel.

0072For example, if the terminal T has switched from the channel C1 to the channel C2, once the detected switchover (25), the access point AP updates (26) a reference table Tb_r to associate the terminal T with the channel C2. This detection is for example obtained when the terminal T, previously present on the channel C1, sends a data frame to the access point AP using C2. The access point receives the data frame and verifies the identity of the transmitting terminal. It then recognizes the AID of the terminal T and then detects that the latter has switched from channel C1 to channel C2. As a result, the access point changes the Tb reference table_r and indicates that the terminal T is no longer associated with the channel C1 but is now associated with the channel C2. We then obtain an updated reference table Tb_ru which is used as a reference when the access point detects a new failover.

0073Furthermore, according to some embodiments, the access point, knowing the availability of all the channels it has in management, can detect the need to switch the terminal T. In this case, it issues (27 ) on channel C1 a command C_B channel change allocated to the terminal T, so that it switches to the C3 channel for example. The C command_B contains the AID of the terminal T, so that the terminal T being placed on C1 understands that this switching command is intended for it.

0074For the sake of simplicity, the processing of a single terminal of the allocation method is shown. In some cases, the same approach can be implemented for a subset of terminals among the set of terminals associated with the same channel managed by the access point. These subsets can be predefined, or defined by the access point at a given time.

0075We present now, in relation to the <figref idref="f0002">figure 3</figref>, the detailed structure of the channel change control, according to a particular embodiment.

5.3 Detailed description of a channel change command

0076We consider, in relation to the <figref idref="f0002">figure 3</figref>, a channel change command, which can be transmitted in the signal transmitted by the access point on a given channel. This command contains the identifier 350 of one or more terminals whose switching is necessary, or information designating this or these terminals (for example in the form of a bitmap field, as specified below).

0077A channel change command contains several bytes, for example:<ul><li>a byte 31 indicating the nature of the command;</li><li>a byte 32 indicating the action of the command;</li><li>five bytes 33 indicating the characteristics of the destination channel of the switchover, for example, the identifier characterizing the structure of the data of the command, the number 332 of bits necessary for the coding of these data, the switching mode 333, the number 334 of the destination channel, the time 335 before the switchover is made, expressed in number of beacons <i>"Beacon"</i>).</li></ul>

0078The channel change control further includes the field 350 to identify the terminal that will switch.

0079According to a first variant, this field 350 contains, for example, two pieces of information, corresponding respectively to a field 34 indicating the length of bits on which the AID of the terminal that is going to change channels is coded, and the field corresponding to the identifier AID of the terminal that will switch, for example coded in binary.

0080According to a second variant, it is considered that there are Z terminals on the same channel, numbered from 0 to Z-1. It is proposed to use a bitmap field 35 of length 34 equal to Z where the Nth bit corresponds to the Nth terminal among the Z terminals present on the same original channel. The bit value indicates whether the terminal is affected by the channel change. It is thus easy to transmit the switching command to a subset of terminals, in a single command.

0081For example, it is considered that there are Z = 5 terminals T1, T2, T3, T4, T5 on the channel C1 and that the AP access point wants to issue a command to switch only M = 2 terminals, by example terminals T2 and T4, on a channel C2. Thus, the field 34 will indicate that the length of the bitmap field is 5, and the bitmap field will take for example the following form:<tables id="tabl0001" num="0001"><table frame="all"><tgroup cols="5" rowsep="0"><colspec colnum="1" colname="col1" colwidth="18mm" /><colspec colnum="2" colname="col2" colwidth="18mm" /><colspec colnum="3" colname="col3" colwidth="18mm" /><colspec colnum="4" colname="col4" colwidth="18mm" /><colspec colnum="5" colname="col5" colwidth="18mm" /><tbody><row rowsep="1"><entry align="center">0</entry><entry align="center">1</entry><entry align="center">0</entry><entry align="center">1</entry><entry align="center">0</entry></row></tbody></tgroup></table></tables>

0082The value 0 corresponds to a holding of the terminals T1, T3 and T5 on the channel C1, and the value 1 corresponds to the switchover of the terminals T2 and T4 on the channel C2.

0083Furthermore, as long as the channel switch has not been performed it is possible to repeat each beacon transmission on the original channel, the channel change command. For example, in the case where the failover is performed only after three tags, that is to say that the field 335 indicates that the channel switching will take place after three tags, we can call back on the original channel within the three beacons the channel change command.

0084For the switchover to be fast, in other words so that directly after the switchover, the terminal can continue its transmission, it is necessary that the terminal receives the beacon of the destination channel almost immediately after the switchover.

0085Traditionally, beacons are sent every 100 ms on a given channel and the channel change command is sent just before the transmission of a beacon on the original channel. Accordingly, if the access point manages asynchronous channels, which is especially contemplated in future standards 802.11ac and 802.11ad, the delay before the flip-flop can retransmit may be long.

0086Furthermore, the field 350 further comprises the identifier 36 of the access point in this channel (BSSID). Thus, according to one embodiment, the identifier of the access point can be, if necessary, modified.

0087We now present, in relation to <figref idref="f0003">Figures 4, 5a to 5c</figref>, an embodiment implementing asynchronous channels to reduce this delay.

5.4 Description of an Embodiment Implementing Asynchronous Channels

5.4.1 First variant implementing asynchronous channels

0088By asynchronous channels, channels with different CSMA-CA access modes from one channel to another are considered, that is to say that the tags (in English " <i>beacons</i> ) Of each channel, for example, carry different information. This information may notably make it possible to know the state of the channel, for example the identity of the access point (BSSID for, in English, "<i>Basic Service Set Identification</i> "), The frequency band, the channel number in this frequency band, the options supported by the PHY / MAC layers.

0089The <figref idref="f0003">figure 4</figref> represents for example two channels C1 and C2 on which the access point sends tags 41 and 42 having different characteristics. In this case, the<figref idref="f0003">figure 4</figref> shows that the tags 41 and 42 are transmitted by the access point with departure times t₀₁ and t₀₂ different.

0090Accordingly, if the access point issues just before the tag 42 a command C_B of channel change on the original channel C2 to the destination channel C1, the terminal that switches to C1 will have to wait a time T_AT to be able to transmit again on channel C1.

0091To reduce this waiting, a synchronization step (40) of the departure times t₀₁ and t₀₂ is proposed, in order to obtain a simultaneous or almost simultaneous emission, beacons 41 and 42.

0092According to another embodiment (not shown) this synchronization can also correspond to applying an identical time interval T separating the transmission of two successive tags 41 and 42, respectively, if these intervals are originally different.

5.4.2 Second variant implementing asynchronous channels offering a quasi-instantaneous switching

0093According to another particular aspect, it is sought to further accelerate the switchover of a terminal from an original channel to a destination channel.

0094To do this, it is anticipated, as illustrated by the <figref idref="f0003">figure 5a</figref>, a transmission by the access point on the original channel C1 of information 51 relating to the state of at least one channel separate from the channel C1, the separate channel corresponding to the destination channel of the switchover. This information 51 relating to the state of the destination channel may correspond to the beacon, or to a part of the beacon (a selection of the useful information), of the destination channel, said beacon being different from the beacon 52 of the channel C1. 'origin.

0095According to a particular aspect represented by <figref idref="f0003">Figures 5b and 5c</figref>, information 51 relating to the state of the destination channel can be directly integrated, by encapsulation, in the tag 52 of the original channel. This gives a single data entity 53 which corresponds, for example according to the<figref idref="f0003">figure 5c</figref>at the insertion of the information 51 relating to the state of the destination channel in the beacon 52 of the channel C1.

5.5 Detailed Description of a Switching Method Implemented in a Terminal According to the Invention

0096We present in relation with the <figref idref="f0004">figure 6</figref>, the switching method implemented in a terminal T according to the invention.

0097Thus, during an association and authentication phase 61 with the access point AP managing at least two communication channels, the terminal T obtains (62) a unique identifier AID. The unique identifier AID is independent of the channel used at a given moment.

0098The terminal T exchanges (63) also an encryption key K with the access point which allows it to authenticate itself. As a result, thanks to this association and authentication phase (61), the identity of the terminal T is shared and known, or available, of all the channels managed by the access point, which allows it to switch from one channel to another without renegotiating this information at the destination channel.

0099Once the association to the access point is made, a failover management phase (64) is activated, during which the terminal can change channels, if necessary. To do this, the terminal determines (65) a destination channel of its failover, among the channels managed by the access point to which it is attached.

0100The invention proposes two variants in the terminal for determining the destination channel of a possible switchover. These two variants are not exclusive of each other, and can therefore, in some embodiments, be implemented one and the other.

0101According to a first variant, the terminal receives a channel change command allocated to said terminal, comprising the identifier AID of the terminal, or information making it possible to locate this terminal (for example the field <i>"Bitmap"</i> already described) and an identifier of the destination channel. It is therefore the access point which, according to this variant, makes the decision of the switchover, for example after analyzing the load on the different channels it manages, to optimize the use of available resources.

0102According to a second variant, the terminal can decide for itself to change the channel. For this, it must receive on the original channel at least one information relating to the state of at least one other channel managed by the access point. This information may in particular be the beacon of this other channel, transmitted for example encapsulated in the beacon of the original channel, as described above.

0103The terminal processes this information and assesses whether another channel, called destination, separate from the original channel, is more suitable for its transmission and therefore decides, of itself, to switch or not to this separate channel that it then determines as its destination channel.

0104According to this particular approach of the invention, the terminal can receive the information relating to several channels managed by the access point to which it is attached, or even all the channels managed by the access point. The processing done by the terminal, certainly more complex, then determines the best channel to perform its transmission.

0105Once this destination channel determined, the terminal switches (66) to the destination channel. Once on the destination channel, the terminal transmits data to the access point, containing its identifier AID. The access point can thus be aware of the actual switchover of the terminal and update a reference table associating the terminal T, identified by its unique identifier AID, to the previously determined destination channel.

5.6 Structure of an access point

0106We present, in relation to the <figref idref="f0004">figure 7a</figref>, the simplified structure of an access point according to the embodiments described above.

0107Such an access point comprises a memory 70 comprising a buffer memory, a processing unit 71, equipped for example with a microprocessor μP, and driven by the computer program 72, implementing the resource allocation method of data transmission according to the invention.

0108At initialization, the code instructions of the computer program 72 are for example loaded into a RAM memory before being executed by the processor of the processing unit 71. The processing unit 71 receives as input signals of a plurality of channels. The microprocessor of the processing unit 71 implements the steps of the data transmission resource allocation method described above, according to the instructions of the computer program 72.

0109For this purpose, the access point comprises, in addition to the buffer memory 70, means for associating and authenticating a terminal with said access point, associating said terminal with a unique identifier in said access point. means for managing said set of channels, implementing a reference table associating with a terminal, identified by its unique identifier, one of said channels, means for detecting a switchover of a terminal attached to said access point, from a first channel, said original channel, to a second channel, said destination channel, and means for updating said reference table to associate the unique identifier of said terminal to said destination channel. These means are controlled by the microprocessor of the processing unit 71.

5.7 Structure of a terminal

0110We finally present, in relation to the <figref idref="f0004">figure 7b</figref>, the simplified structure of a terminal according to the embodiments described above.

0111Such a terminal comprises a memory 75 comprising a buffer memory, a processing unit 72, equipped for example with a microprocessor μP, and driven by the computer program 73, implementing the switching method according to the invention.

0112At initialization, the code instructions of the computer program 73 are for example loaded into a RAM before being executed by the processor of the processing unit 74. The processing unit 74 receives as input a piece of information. an access point managing at least two communication channels. The microprocessor of the processing unit 74 implements the steps of the selection method described above, according to the instructions of the computer program 73.

0113For this, the access point comprises, in addition to the buffer memory 75, association and authentication means with said access point, comprising means for obtaining a unique identifier in said access point, and failover management means implementing means for determining a destination channel managed by said access point, by processing at least one piece of information received from said access point and means for switching the transmission data from said original channel to said determined destination channel, without modifying said identifier. These means are controlled by the microprocessor of the processing unit 74.