GB2508158A

GB2508158A - Synchronising a transmitter and receiver in a CSMA ad-hoc network

Info

Publication number: GB2508158A
Application number: GB201220931A
Authority: GB
Inventors: Romain Guignard; Pascal Viger; Patrice Nezou
Original assignee: Canon Inc
Current assignee: Canon Inc
Priority date: 2012-11-21
Filing date: 2012-11-21
Publication date: 2014-05-28
Anticipated expiration: 2032-11-21
Also published as: GB201220931D0; GB2508158B

Abstract

A transmitter and receiver in an ad-hoc network synchronise an application event, e.g. Start Of Frame 300 of a video source. The transmitter and receivers of the network share knowledge of some future expected network events, e.g. future medium access events based on carrier sense multiple access (CSMA) backoff values, and the transmitter chooses one of the future network events 322 (T4) as a reference. It transmits synchronization information 321 comprising an index of this chosen event and a time interval 311 related to the time difference between the application event at the transmitter and the reference network event. The receiver computes the time to trigger a corresponding receiver application event 301 from the received information and knowledge of a synchronization delay 330. Accordingly, the synchronisation of diverse applications is managed without disturbance in an unstable network. This generic synchronization protocol is also bandwidth efficient.

Description

METHOD FOR SYNCHRONIZING A TRANSMITTER AND A RECEIVER IN

A CSMA AD-HOC NETWORK AND CORRESPONDING DEVICES

The present invention concerns a method and a device for synchronizing a transmitter and a receiver in a CSMA ad-hoc network.

In many applications with transmitter/receiver communication, application synchronization is necessary between the source and the receiver. For example, in a live video streaming application between a source and one or more receivers, the applicative data havc to bc rendcrcd to thc application with thc same temporal characteristics, for instance thc samc framc ratc, as in thc original data strcam. Thcreby, thc start of cach frame has to be rendered with the same cadence as the source. However, each node has a local oscillator from which is generated the application clock. Although, the local oscillators of each of the nodes have the same nominal value, the imperfections of those oscillators lead to a drift between the application clocks for the different nodes, between the source node and the receiver node or between a plurality of receivers. This phenomenon can lead to quality of service disturbance concerning the application of live video streaming but also some overflow or underflow in the memories of the different stages of the system.

The invention lies in the context of a communication network sharing a common transmission medium. In such networks, network events, for instance actual transmission of a data packet, constitute temporal markers relative to the network clock used by all the communication nodes connected to the network. An example of such a networks is a LAN (Local Area Network) sharing the same transmission medium, being wired of wireless, like Ethernet, Wi-Fi or IEEE 1394. There are many others.

In other systems, the synchronization could have a different impact. For instance, a video acquisition system with several acquisition devices, namely cameras, is used to build a 3D vision or bullet time effect, a kind of slow motion with different points of view. For those applications, the synchronization impacts the precision or coherency of the different views and thus results in disturbance in the 3D reproduction of the subject shot by the cameras. Audio acquisition systems used for source location also need accurate synchronization between the different sensors. Indeed, a phase difference between the sensors results in a mismatch of the position of the subject to locate. The aforementioned applications require synchronization between several receivers in order to ensure a simultaneous start cycle for all the receivers.

In most cases, the application synchronization depends on the MAC layer synchronization from which is recovered the timestamp used at the application layer.

Real-time Transport Protocol (RTP) or MPEG-Transport Stream may be cited as examples of this. Many protocols are available to synchronize the MAC layer. For instance, the well-known protocols 1EEE1588: Precision Time Protocol (PTP) or Network Time Protocol (NTP) and also the synchronization protocol implemented in IEEE 1394 in its wired or wireless versions. Those protocols often require there to be a master node (e.g. a base station) which shares its local clock to the other nodes in order to achieve the synchronization. Tn 1EEE1394, the master node sends its local timestamp periodically through a register in a beacon frame. This timestamp is compared in the receivers with their local timestamp so as to compute and correct their drift. The 1EEE1588 protocol uses a point-to-point master/slave synchronization mechanism with two successive round-trip messages. However, those protocols are not well adapted to an ad-hoc environment wherein the condition of transmission and the network topography are unstable i.e. the access to the medium and the bandwidth are not guaranteed and it is difficult to select a master. Moreover, the implementation of the MAC layer and application layer synchronization may be demanding in some systems and involve frequently and periodically sending information relating to the local clock which is very bandwidth consuming.

In this document, ad-hoc network means a network where all the connected nodes share the same status. There is no master-slave scheme implemented at the network layer.

An ideal synchronization mechanism should provide a solution to be able to manage the synchronization of diverse applications without disturbance in an unstable environment. This generic synchronization protocol should also be bandwidth efficient i.e. send only the necessary information; but this is not achievable with state of the art technologies.

The present invention has been devised to address one or more of the foregoing concerns.

According to a first aspect of the invention there is provided a method of sending synchronization information relating to an application event by a transmitter, the transmitter being connected to at least one receiver through a communication network, comprising detecting a transmitter application event; choosing a reference network event among next expected network events, the knowledge of at least some of the next expected network events being shared with the at least one receiver; determining a time interval related to the received transmitter application event and the time of occurrence of the reference network event; and sending to the at least one receiver synchronization information including an index of the chosen reference network event and the determined time interval. Accordingly, the synchronization of diverse applications is managed without disturbance in an unstable environment. This generic synchronization protocol is also bandwidth efficient.

In an embodiment, the chosen reference network event is the last expected network events occurring before the expected time of occurrence of a corresponding receiver application event which corresponds to the time of occurrence of the transmitter application event plus a given synchronization delay. Accordingly, the drift in the receiver is minimized.

In an embodiment, the method further comprises estimating the time of occurrence of the reference network event, the network access being controlled by a CSMAJCA scheme, based on a shared knowledge of the backoff values used by the transmitter and the at least one receiver. Accordingly, the proposed method is adapted to a CSM A network.

In an embodiment, the method further comprises estimating a new the time of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the sending of the synchronization information; and updating the determined time interval accordingly. Accordingly, the method increase the precision in case of unexpected network event.

In an embodiment, the time interval corresponds to the time difference between the time of occurrence of the reference network event and the time of occurrence of the transmittcr application event. Accordingly, the computations at the transmitter arc kept minimal.

In an embodiment, the time interval corresponds to the time difference between the time of occurrence of the reference network event and the expected time of occurrence of the corresponding receiver application event, the expected time of occurrence of the corresponding receiver application event corresponding to the time of occurrence of the transmitter application event plus a given synchronization delay.

Accordingly, the computations at the receiver are kept minimal.

According to another aspect of the invention there is provided a method of triggering an application event by a receiver connected to a transmitter through a communication network, thc knowledge of at least some of thc next cxpcctcd network cvcnts bcing sharcd with thc transmitter and potentially with other rcccivcrs connected to the transmitter wherein the method comprises receiving synchronization information including a time interval and an index of a reference network event among the next expected network events; determining a waiting time to wait after the time of occurrence of the reference network event based on the received time interval; and triggering the application event at the end of the waiting time after the occurrence of the reference event.

In an embodiment, the received time interval being the time interval between the time of occurrence of a transmitter application event and the time of occurrence of the reference network event, the waiting time is determined by taking a given synchronization delay minus the received time interval.

In an embodiment, the received time interval being the time interval between the time of occurrence of the reference network event and the expected time of occurrence of the corresponding receiver application event which time of occurrence corresponds to the time of occurrence of the transmitter application event plus a given synchronization delay, the waiting time is determined by taking the value of the received time interval.

In an embodiment, the method ifirther comprises estimating the time of occurrence of the reference network event, the network access being controlled by a CSMAICA scheme, based on a shared knowledge of the backoff values used by the transmitter and the receivers participating in the method.

In an embodiment, the method further comprises determining the difference between the actual time of occurrence of the reference network event and the estimated timc of occurrence of the reference network event; and adjusting the waiting time accordingly. Accordingly, the precision in the triggering of the receiver application event is increased.

In an embodiment, the method further comprises estimating anew the time of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the actual time of occurrence of the reference network event; and updating the waiting time accordingly.

According to another aspect of the invention there is provided a method for synchronizing an application event between a transmitter connected to at least one receiver through a communication network, wherein the method comprises sending synchronization information relating to the application event by the transmitter according to the invention; and triggering an application event by the at least one receiver, based on said synchronization information, according to the invention.

According to another aspect of the invention there is provided a transmitter for sending synchronization information relating to an application cvcnt, thc transmitter being conncctcd to at least one receiver through a communication network, wherein the transmitter comprises a detecting module for detecting a transmitter application event; a choosing module for choosing a reference network event among next expected network events, the knowledge of at least some of the next expected network events being shared with the at least one receiver; a determining module for determining a time interval related to the received transmitter application event and the lime of occurrence of the reference network event; and a sending module for sending to the at least one receiver synchronization information including an index of the chosen reference network event and the determined time interval.

In an embodiment, wherein the choosing module is adapted to choose the reference network event is the last expected network events occurring before the expected time of occurrence of a corresponding receiver application event which corresponds to the time of occurrence of the transmitter application event plus a given synchronization delay.

In an embodiment, wherein the transmitter ifirther comprises an estimating module for estimating the time of occurrence of the reference network event, the network access being controlled by a CSMA/CA scheme, based on a shared knowledge of the backoff values used by the transmitter and the at least one receiver.

In an embodiment, the estimating module is adapted for estimating a new the time of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the sending of the synchronization information; and wherein the transmitter further comprises an updating module for updating the determined time interval accordingly.

In an embodiment, the time interval corresponds to the time difference between the time of occurrence of the reference network event and the time of occurrence of the transmitter application event.

According to another aspect of the invention there is provided a receiver adapted for triggering an application event, the receiver being connected to a transmitter through a communication network, thc knowlcdgc of at least some of thc ncxt expected network cvcnts bcing sharcd with thc transmitter and potcntially with other rcccivcrs conncctcd to the transmitter wherein the receiver comprises a receiving module for receiving synchronization information including a time interval and an index of a reference network event among the next expected network events; a determining module for determining a waiting time to wait after the time of occurrence of the reference network event based on the received time interval; and a triggering module for triggering the application event at the end of the waiting time after the occurrence of the reference event.

In an embodiment, the received time interval being the time interval between the time of occurrence of a transmitter application event and the time of occurrence of the reference network cvcnt, the waiting time is determined by taking a given synchronization dclay minus the received time intcrval.

In an embodiment, the receiver further comprises an estimating module for estimating the time of occurrence of the reference network event, the network access being controlled by a CSMA/CA scheme, based on a shared knowledge of the backoff values used by the transmitter and the receivers participating in the method.

In an embodiment, the receiver further comprises a determining module for determining the difference between the actual time of occurrence of the reference network event and the estimated time of occurrence of the reference network event; and an adjusting module for adjusting the waiting time accordingly.

In an embodiment, the estimating module is adapted for estimating anew the timc of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the actual time of occurrence of the reference network event; and the receiver further comprises an updating module for updating the waiting time accordingly.

According to another aspect of the invention there is provided a transmission system for synchronizing an application event between a transmitter connected to at least one receiver through a communication network, wherein the transmission system comprises a transmitter for sending synchronization information relating to the application event by the transmitter according to the invention; and at least a receiver for triggering an application event, based on said synchronization information, according to the invention.

According to another aspect of the invention there is provided a computer program product for a programmable apparatus, the computer program product comprising a sequence of instructions for implementing a method according to the invention, when loaded into and executed by the programmable apparatus.

According to another aspect of the invention there is provided a computer-readable storage medium storing instructions of a computer program for implementing a method according to the invention.

At least part of the methods according to the invention may be computer implcmcntcd. Accordingly, thc prcscnt invdntion may takc thc form of an cntircly hardwarc cmbodimcnt, an cntircly softwarc embodimcnt (including firmware, rcsidcnt software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a circuit', "module" or "system".

Furthcrmorc, thc prcscnt invcntion may talc c thc form of a computcr program product embodied in any tangible medium of expression having computer usable program code embodied in the medium.

Since thc present invcntion can bc implcmcnted in software, thc prcscnt invention can be embodied as computer readable code for provision to a programmable apparatus on any suitable carrier medium. A tangible carrier medium may comprise a storage medium such as a floppy disk, a CD-ROM, a hard disk drive, a magnetic tape device or a solid state memory dcvicc and thc like. A transicnt carrier medium may include a signal such as an dectrical signal, an electronic signa', an optical signaL an acoustic signal, a magnetic signal or an electromagnetic signal, e.g. a microwave or RF signal.

Embodiments of the invention will now be described, by way of example only, and with reference to the following drawings in which: Figure 1 describes a video streaming system in which the invention may be implemented; Figure 2 is a block diagram illustrating a schematic configuration of a communication apparatus representing a transmitter or a receiver adapted to incorporate the invention; Figure 3 is a representation of the different steps to carry out at the transmitter according to an embodiment of the invention; Figure 4 is a representation of the different steps to carry out at the receiver according to an embodiment of the invention; Figure 5 illustrates the behaviour of the receiver when an extra node accesses to the medium according to an embodiment of the invention; Figure 6 illustrates the main steps of the method at the transmitter to implement an embodiment of the invention; Figure 7 iHustrates the main steps of the method at the receiver to implement an embodiment of the invention.

Figure 1 describes a video streaming system in which the invention can be implemented. This distribution system is used to display the content of a video source on a screen. This kind of architecture is an example of a communication system where the present invention can be implemented. It is based on a communication network 100 sharing common network events. This communication network uses a shared transmission network. Thc communication between all the communicating nodes is controlled using a multiple access method. In the example, the network is controlled based on a Carrier sense multiple access with collision avoidance (CSMA/CA). This is a network multiple access method in which carrier sensing is used, but nodes attempt to avoid collisions by transmitting only when the channel is sensed to be idle'. Other multiple access method may be used, for example time division multiple access (TDMA) or others.

The system comprises a sending node 111, here below also called a sending device, comprising a first end connected to the digital network 100 and a second end connected to a video source 110, for example a DVD player, generating raw audiovisual applications data. Note that the video source can also generatc encoded vidco data. In some embodiment, the video source 110 and thc sending node 111 may bc grouped in a single device.

The system comprises a receiving node 120, here below also called a receiving device, comprising a first end connected to the digital network 100 and a second end connected to a video display device or screen 121. In some embodiment, the receiving device 120 and the video display device 121 may be grouped in a single device.

The system also comprises a set of nodes 101, 102, 103, 104 and 105 which exchange data through the digital network 100. These nodes may be also receiving devices participating to the invention or legacy communication devices using the same network for their own purpose.

The network generates some network events. These network events when they occur represent temporal markers that may be used as synchronization point between nodes. They may be generated, for example, by the start of the transmission by a node on the network. This is the case in the detailed example of embodiment of the invention.

According to a first aspect of the invention, all the nodes participating to the invcntion share the knowledge of the expected network events. This knowledge may be limited to the next few network events. As explained in detail below, this knowledge may be limited to a single event which will be then chosen as the reference network event and the network event used for the transmission of the synchronization information. In the preferred embodiment, in order to allow choosing the best reference network event, this knowledge covers a time period exceeding the latency of the system.

Their temporal location, if not known precisely should at least be predicted by the different nodes. For example, in a TDMA network, the transmission time by the different nodes occur in reserved time-slots. Therefore this knowledge is an inherent feature of the network. In some other kind of networks this may not be the ease.

In a typical CSMA/CA network, a node having data to transmit senses the medium to test if it is idle or occupied. If the medium is idle, the node can begin a transmission. It takes the medium and transmits its data. The transmission time, while being bounded depends on the quantity of data to be transmitted and is not known a priori. If the medium is occupied, the node randomly chooses a counter value, called a baekoff, and begins to count down. It makes a new transmission attempt when this counter reaches zero. Moreover, the countdown is interrupted when another node takes the medium for transmission.

According to this mechanism a particular node is unable to know, or even to predict, the transmission time by other nodes. In such a network, the common knowledge of the expected network events may be achieved by the following mechanism implemented by all the nodes participating in the invention. To be able to determine the time when another node has an opportunity to transmit the nodes must share their backoff values. Many ways to achieve the sharing of backoff values may be contemplated. These values may be broadcast by the nodes to the others. Another way consists in having all the nodes share their pseudo random generator seeds. By doing so, all the nodes may reproduce the pseudo random generation of the backoff values of other nodes at no transmission cost. The next transmission opportunity for a node may be estimated based on the knowledge of its backoff value. As already said, the countdown is interrupted by the actual transmission occurring during this countdown.

The waiting time associated with a backoff value is not known in advance but may be estimated between two bounds given by the following functions. The lower bound is given by the backoff value multiplied by the duration of a countdown step. It corresponds to the case where no transmission at all occurs during the countdown. The upper bound is given by the countdown value multiplied by the maximum transmission time. It corresponds to transmission occurring at each count down, each transmission lasting the maximum transmission time. The actual time is somewhere in between depending on the network load. This time may be estimated by taking the lower bound, the upper bound or a weighted combination of the two. The weight is a parameter representing the expected load of the network. Whatever the method actually chosen to estimate the time of occurrence of the next few expected network events, it has to be the same for all the nodes participating in the invention. In this way, all the nodes share the same knowledge of the next few network events.

The methods herein described as pertaining to the invention are typically implemented within the sending node 111 and the receiving node 120.

Figure 2 is a block diagram illustrating a schematic configuration of a communication apparatus representing a transmitting device or a receiving device adapted to incorporate the invention. Reference numeral 202 is a RAM which functions as a main memory, a work area, etc., of CPU 201, and the memory capacity thereof can be expanded by an optional RAIVI connected to an expansion port (not illustrated). CPU 201 is capable of executing instructions from program ROM 203 on powering on of the communicating apparatus. After the powering on, CPU 201 is capable of executing instructions from main mcmory 202 rclating to a softwarc application aflcr those instructions have bccn oadcd from thc program RUM 203 or thc hard-disc (HD) 206 for example. Such a software application, when executed by the CPU 201, causes the steps of the flowcharts shown in Figures 6 and/or 7 to be performed.

Reference numeral 204 represents the network interfaces that can be a single network interface, or composed of a set of different network interfaces (for instance wired and wireless interfaces, or different ldnds of wired or wireless interfaces). Data packets arc written to the nctwork intcrfacc for transmission or read from thc network intcrfacc for rcccption under the control of the software application running in thc CPU 201. Reference numeral 205 represents a user interface to display information to, and/or receive inputs from, a user.

1/0 moddc 207 reprcscnts a moddc abic to rcceivc or send data from/to external devices like video source or display).

Figure 3 illustrates the mechanism set up at the transmitter for the synchronization in an embodiment of the invention. This embodiment is described in the context of a CSMAJCA network but the person skilled in the art would understand that the mechanism would apply in other kinds of networks, for example TDMA.

The first timelinc represents the application evdilts occurring at the application layer in the transmitter. An application event 300 occurs, namely a Start of Frame ill the example, which must be synchronized with a corresponding application event 301 on thc rccciver side. Thc corrcsponding application cvdnt will occur at the rccciver after a latency delay 330, also called synchronization delay, which will be used to achieve the synchronization. The application events occur in a time based on the application clock.

The second timeline represents the network events occurring at the network layer. These network events 320, 321, 322 and 323 represent the start of transmission of the different nodes connected to the network. These network events occur in a time based on the network clock.

The invention aims at synchronizing the transmitter application event 300 with the corresponding receiver application event 301 each occurring based on its own application clock. The synchronization delay reflecting the latency is a system parameter known by the nodes participating in the invention. The network events are used as a relay to synchronize the application clock of the transmitter and the application clock of the receiver.

Thc application cvcnt 300 must bc timc stamped using the nctwork clock. To do so, each time a network event occurs, a timer is triggered at the network layer. This timer is stopped when an application event occurs allowing it to be time stamped with the network time. In the example, the timer is triggered on the occurrence of network event 320 at T1. The timer is stopped on the occurrence of the application event 300 at T2. The time At referenced 310 and equal to T2 -T1 allows the application event 300 to be time stamped according to the network clock.

Next a particular network event must be chosen to serve as a rcfercnce for thc synchronization. In theory all the network events may be chosen to meet this need.

Actually some are more advantageous than others. As the chosen network event acts as a synchronization point, the less the application clock at the receiver has time to drift after the occurrence of this network event and before the triggering of the corresponding application event the better, since this will minimize the application clock drift at the receiver. Therefore the last network event occurring before the estimated application event at the receiver is the best choice to act as the reference network event. Namely, on Figure 3, the network event 322 is chosen by the transmitter as the reference network event. This choice occurs at T2 before the occurrence of the network event 322 which occurs at T4. This choice is possible thanks to the a priori knowledge of the next few network events shared by all the nodes participating in the invention.

Next, the time interval between the occurrence of the application event 300 and the chosen reference network event 322 must be determined. This time interval referenced 311 may be precisely known depending of the kind of network used. For example, in a TDMA, the time T4 corresponds to a given time-slot in the TDMA grid and is perfectly known with respect to T1. Therefore the computation of At1 is straightforward. In our example based on a CSMAICA network, the precise occurrence timc T4 of the reference nctwork event 322 is not known precisely and must be estimated. As already described in the foregoing, various estimation methods exist. One of these estimation methods is then used to obtain an estimation of T4. M is then determined based on this estimation.

The chosen reference network event and the corresponding time interval 311 between the occurrence of the application event and the reference network event need to be transmitted to the receiver. This transmission must occur before the reference network event, meaning between T2 and T4. According to a first embodiment, the latency is chosen large enough to ensure that an opportunity to transmit occurs for the transmitter between T2 and T4. This opportunity is then used by the transmitter to transmit the data. But this may lead to an undesirably large latency. Advantageously a media sharing mechanism is implemented between the nodes participating in the invention. According to this mechanism, a time slot is reserved during all transmissions by any node to allow signalling information to be exchanged with another node. This allows a small out of band transmission channel to be implemented which can be used by the transmitter to transmit the time interval M51 to the receiver. By using this mechanism, the transmitter is able to take advantage of any network event occurring after the application event at 2 and before the reference network event at T4 to transmit the synchronization information. For example, in Figure 3, the transmitter will use the network event 321, whichever the node taking the medium, to transmit the synchronization information out of band to the receiver. This mechanism allows the latency to be decreased by reducing the synchronization delay by using a network event to transmit the synchronization information. In a particular embodiment, the synchronization information can be transmitted in the time slot of the reference network event itself The synchronization information includes the time interval between the application event and the reference network event, represented by its index, or at least an estimation of the latter. The indication of the reference network event may be provided in an absolute manner or relatively to the transmission event.

Figure 4 illustrates the mechanism set up at the receiver for the synchronization in an embodiment of the invention.

At time T3, the transmitter transmits the synchronization information. This transmission constitutes the network event 421. It may be that the transmitter has taken the medium for transmission. It may also be that another node takes the medium and implements the mechanism that reserves a time slot during the transmission to allow the transmitter to transmit the synchronization information. The receiver, then receives the synchronization information including the time interval t1 and an index of the reference network event 422 at the time T3 -This information is used to determine the time T4 of the expected occurrence of the reference network cvcnt. The determination may be straightforward in a TDMA nctwork or be based on an cstimation function similarly to thc estimation made by the transmitter. This is what happens in a CSMAJCA network for example. The estimation function used to estimate the time of occurrence of the reference network event 422 must be the same estimation function as the one used by the transmitter. Accordingly, the same time T4 is estimated by both the transmitter and the receiver.

The time T4 having been estimated, the knowledge of t1 allows determination of the time T'2 corresponding to the application event at the transmitter by using the formula: T'2 = T4 -The synchronization delay 430 being a system parameter also known at the receiver, the waiting delay Atrt between the time T4 of occurrence of the reference network and the time T'5 when the receiver application event must be triggered can be determined using the formula: At1 = Synchronization_delay -At51; When the reference network event actually happens, there are two possibilities.

A first of the possibilities is that it happcns at the expected time T4. A countdown is then triggered with the value Atri to wait for the time T'5. A second of the possibilities is that the reference network event occurs at a time T'4 slightly different to the expected time T4. Then the difference between the expected and the actual time of occurrence of the reference network is used to adapt the value of the waiting delay Ati. A countdown is then triggered with the adapted value Atrt to wait for the time T'5.

When the countdown reaches zero, at time T'5 the receiver application event 401 is triggered. Altematively, the counter may be initialized to zero and incremented until it reaches the value At1. Thanks to the invention this happens exactly with a time differcncc corresponding to the synchronization delay with respect to the transmitter application event.

Figure 5 illustrates a ease where a legacy node not participating in the invention takes the medium instead of the expected node between the estimation of the time of occurrence of the reference network event and the actual time of occurrence of the reference network event. In the example, this happens at the expected time of occurrence of the reference network event. This node generates an unexpected network event. This unexpected network event does not belong to the next few expected network events the knowledge of which is shared by the nodes participating in the invention. The first steps of the described mechanism are similar to the one described relatively to Figure 4.

The receiver expects the reference network event 522 referred in the received synchronization information to occur at T4. Instead, at T4 a legacy node not participating in the invention takes the medium and starts a transmission. This constitutes the network event 523. In a first embodiment, the receiver uses this network event 523 which occurs at, or near, the expected time 1'4 to serve as the expected network event. In that case, the next operations are similar to the ones described relatively to Figure 4.

In some embodiments, the fact that all the nodes participating in the invention are aware of such a network event triggered by a legacy node may not be guaranteed. It may happen that a receiver participating in the invention and which needs to be synchronized does not see the network event 523. Such a node is still waiting for the expected network event 522. Therelbre, in an alternative embodiment, the reference network cvcnt 522 is kcpt as thc reference for thc calculation of thc application event in the receiver.

According to this embodiment, at T4 when the occurrence of the reference network event is expected, the receiver launches a counter to track the waiting delay.

The occurrence of the network event 522 is then expected at 1'4 after a time shift 513 corresponding to 7 -T4. When the receiver detects that the medium has been granted to a legacy node instead of the expected node, a new prediction of the occurrence of the reference network event is made taking into account the medium access shift 513. The waiting time Ist,. is modified accordingly leading to a new value A't referenced 512'.

The medium access shift time may be known thanks to a dedicated message like, fbr instance, the message NAV in standard 802.11 or it may be estimated. The estimation may be based on the maximum transmission time. The synchronization may then be done as described relatively to Figure 4 with the new estimation T of the occurrence of the reference network event and the modified waiting time A'trl.

In some cases, the new estimation of the occurrence of the reference network event gives a result that occurs after the expected application event T. Namely, the adapted waiting time A't gets a negative value. In this case, at thc time of the occurrence of the reference network event it will be too late for triggering the receiver application event 501. In this case, the counter triggered at T4 is not stopped and the receiver application event is triggered using the initial value of the waiting time St,.1.

This means that the time for triggering the receiver application event is based on the initial estimation of T4 instead of the actual occurrence of the reference network event.

It may lead to a lack of precision in the triggering of the receiver application event 501.

When the reference network event actually occurs, after the triggering of the receiver application event 501, it may be used to compute the potential error made in the triggering of the receiver application event 501 a posteriori. For instance, the potential error may be determined by launching a counter at the triggering of the receiver application event 501 that is stopped on the occurrence of the reference network event at T1. The value of this counter is compared to the negative value of the adapted waiting time 512' to determine the potential error.

The knowledge of this error may be used by the application depending on the kind of service. For example, a correction may be applied to the next start of frame for a video streaming service. This illustrates the robustness of the synchronization mechanism against unpredictable medium access shifting which happens quite often in a. wireless environment like standa.rd 802.11. At the transmitter such a medium access shifting occurring for instance at T3 is corrected before sending the synchronization information. A new estimation of the time T4 of occurrence of the reference network event is made. The time interval atç1 referenced 311 in Figure 3 is updated accordingly before being sent to the receiver.

Figure 6 illustrates the main steps of the method at the transmitter to implement an embodiment of the invention. This process takes place at the transmitter. In a first step 600, the network layer receives an application event from the application layer.

This occurrence is time stamped with the network time.

In a step 610, the transmitter determines a reference network event to be used as a reference for the synchronization of the received transmitter application event and a corresponding receiver application event. Knowing the next few network events to occur, this reference network event is chosen based on the time of occurrence of the receiver application event which must occur after a synchronization delay reflecting the latency of the system. Advantageously, the last network event occurring before the determined time of occurrence of the receiver application event is chosen as the reference network event.

In a step 620, the time of occurrence of this reference network event is determined. For instance, it may be estimated in a CSMAICA network using an estimation function based on the shared knowledge of the backoff of the different nodes participating in the invention. Once this time of occurrence has been determined, the time interval between the occurrence of the transmitter application event and this determined time of occurrence of the reference network M51 is determined.

Alternatively, the waiting time atri may bc directly dctcrmined at the transmitter to be transmitted to the receiver.

In a step 630, the transmitter waits for the next transmission opportunity, meaning a network event where a node participating in the invention takes the medium for transmission. If the medium is available, thanks to the mechanism allowing the transmitter to transmit out of band during the transmission time granted to another node, the transmitter transmits its synchronization information to the receivers in a step 640.

In the described embodiment, this information consists in the index of the chosen reference network event and thc time difference between the transmitter application event and the estimated time of occurrence of this reference network event.

If the medium is not available, usually due to a legacy node taking the medium, a new estimated time of occurrence of the chosen reference network event is determined. The time of occurrence of the reference network event is estimated anew when a node other than the transmitter and the at least one receiver generates a network event between the previous estimation and the sending of the synchronization information. The time intenal is updated accordingly. This is done in the step 650, before a new transmission attempt. Advantageously, if the new time of occurrence determined for the reference network event happens to occur after the expected receiver application event, a new reference network event maybe determined.

Figure 7 illustrates the main steps of the method at the receiver to implement an embodiment of the invention. In a step 700, the receiver detects a new network event, typically a new transmission start. Tn a step 710, the receiver checks if the network event is an expected network event. As already seen, the nodes participating in the invention share a common knowledge of the next few network events. These next few network events concern the network event generated by the nodes participating in the invention. When, for instance, a legacy node not participating in the invention takes the medium, it constitutes an unexpected network event which does not fit within the shared knowledge of expected events.

In case of an unexpected network event and if a synchronization process is running an estimation of the time of occurrence of a reference network is made. This unexpected network event disrupts the estimation by introducing an unexpected transmission time. Therefore, in a step 711, the estimation of the time of occurrence of the reference network event and the associated waiting delay to trigger the receiver application event is updated to take into account the unexpected transmission.

In case, the network event is an expected network event, it may be a transmission used for the transmission of synchronization information from the transmitter. This is tested in step 720.

If this is the case, the receiver receives an index of the chosen reference network event and also receives the time interval between the time of occurrence of the transmitter application event and this reference network event in the step 721.

Using the shared common knowledge of the next few events, the reference network event is then determined by the receiver. In a step 722, its time of occurrence is estimated using the same estimation function as was used within the transmitter.

In a step 723, the waiting delay between the estimated time of occunence of the reference network event and the receiver application event is determined based on the received synchronization information as detailed relatively to Figure 4 and Figure 5.

In case the expected network event detected does not carry synchronization information, it is tested in step 730 if this is a reference network event. If the time of occurrence of this reference network event is not the estimated time, the determined waiting delay is adjusted accordingly. Then in a step 731, a counter is triggered to count this waiting delay.

At the end of this waiting delay, in a step 732, the receiver application event is triggered and notified to the application layer.

When thc current dctectcd network cvcnt is has bcen processed, thc mcthod returns to step 700 to wait for the next network event.

According to this method synchronization between application events is achieved at the application layer between a transmitter and a receiver. Medium access jitter is handled with no extra message needed and no impact on the synchronization.

The MAC layer synchronization mechanisms are not needed. The occurrence of the receiver application event is controlled directly by the corresponding occurrence at the transmitter, there is no need for drift correction. The drift is bounded and the method has a very low overhead. This mechanism may also be used to start several devices simultaneously, like a multi shooting camera system or an audio/video acquisition system.

Although the present invention has been described hereinabove with reference to specific embodiments, the present invention is not limited to the specific embodiments, and modifications will be apparent to a person skilled in the art which lie within the scope of the present invention.

Many further modifications and variations will suggest themselves to those versed in the art upon making reference to the foregoing illustrative embodiments, which are given by way of example only and which are not intended to limit the scope of the invention, that being determined solely by the appended claims. In particular the different feaIires from different embodiments may be interchanged, where appropriate.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that different features are recited in mutually different dependent claims does not indicate that a combination of these features cannot be advantageously used.

Claims

CLAIMS1. A mcthod of sending synchronization information rdating to an application event by a transmitter, the transmitter being connected to at least one receiver through a communication network, wherein the method comprises: -detecting a transmitter application event; -choosing a reference network event among next expected network events, the knowlcdge of at icast some of thc next expeetcd network events being shared with the at least one receiver; -determining a time interval related to the received transmitter application event and the time of occurrence of the reference network event; and -sending to the at least one receiver synchronization information including an index of thc chosen reference network cvcnt and the determined time interval.
2. A method according to claim I, wherein the chosen reference network event is the ast expected network events occurring before the expected time of occurrence of a corresponding receiver application event which corresponds to thc time of occurrence of thc transmitter application event plus a given synchronization delay.
3. A method according to claim I or 2, wherein the method further comprises -estimating the time of occurrence of the reference network event, the network access being controlled by a CSMAICA scheme, based on a shared knowledge of the backoff values used by the transmitter and the at least one receiver.
4. A method according to claim 3, wherein the method further comprises -estimating a new the time of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the sending of the synchronization information; and -updating the determined time interval accordingly.
5. A method according to any one of claims I to 4, wherein the time interval corresponds to the time difference between the time of occurrence of the reference network event and the time of occurrence of the transmitter application event.
6. A method according to any one of claims 1 to 4, wherein the time interval corresponds to the time difference between the time of occurrence of the reference network event and the expected time of occurrence of the corresponding receiver application event, the expected time of occurrence of the corrcsponding rccciver application event corresponding to the timc of occurrcncc of thc transmitter application event plus a given synchronization delay.
7. A method of triggering an application event by a receiver connected to a transmitter through a communication network, the knowledge of at least some of the next expected network events being shared with the transmitter and potentially with other receivers connected to the transmitter wherein the method comprises: -receiving synchronization information including a time interval and an index of a reference network event among the next expected network events; -determining a waiting time to wait after the time of occurrence of the reference network event based on the received time interval; and -triggering the application event at the end of the waiting time after the occurrence of the reference event.
8. A method according to claim 7, wherein, the received time interval being the time interval between the time of occurrence of a transmitter application event and the time of occurrence of the reference network event, the waiting time is determined by taking a given synchronization delay minus the received time interval.
9. A method according to claim 7, wherein, the received time interval being the time interval between the time of occurrence of the reference network event and the expected time of occurrence of the corresponding receiver application event which time of occurrence corresponds to the time of occurrence of the transmitter application event plus a given synchronization delay, the waiting time is determined by taking the value of the received time interval.
10. A method according to any one of claims 7 to 9, wherein the method further comprises: -estimating the time of occurrence of the reference network event, the network access being controlled by a CSMAJCA scheme, based on a shared knowledge of the backoff values used by the transmitter and the receivers participating in the method.
11. A method according to claim 10, wherein the method further comprises: -determining the difference between the actual time of occurrence of the reference network event and the estimated time of occurrence of the reference network event; and -adjusting the waiting time accordingly.
12. A method according to claim 10 or 11, wherein the method further comprises: -estimating anew the time of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the actual time of occurrence of the reference network event and -updating the waiting time accordingly.
13. A method for synchronizing an application event between a transmitter connected to at least one receiver through a communication network, wherein the method comprises: -sending synchronization information relating to the application event by the transmitter according to any one of claims 1 to 6; and -triggering an application event by the at least one receiver, based on said synchronization information, according to any one of claims 7 to 12.
14. A transmitter for sending synchronization information relating to an application event, the transmitter being connected to at least one receiver through a communication network, whcrein the transmitter compriscs: -a detecting module for detecting a transmitter application event; -a choosing module for choosing a reference network event among next expected network events, the knowledge of at least some of the next expected network events being shared with the at least one receiver; -a determining module for determining a time interval related to the received transmitter application event and the time of occurrence of the reference network event; and -a sending module for sending to the at least one receiver synchronization information including an index of the chosen reference network event and the determined time interval.
15. A transmitter according to claim 14, wherein the choosing module is adapted to choose the reference network event is the last expected network events occurring before the expected time of occurrence of a corresponding receiver application event which corresponds to the time of occurrence of the transmitter application event plus a given synchronization delay.
16. A transmitter according to claim 14 or 15, wherein the transmitter ifirther comprises -an estimating module for estimating the time of occurrence of the reference network event, the network access being controlled by a CSMA/CA scheme, based on a shared knowledge of the backoff values used by the transmitter and the at least one receiver.
17. A transmitter according to claim 16, wherein the estimating module is adapted for estimating a new the time of occurrence of the reference network event when a node generates an unexpected network event between the previous estimation and the sending of the synchronization information; and wherein the transmitter further comprises -an updating module for updating the determined time interval accordingly.
18. A transmitter according to any one of claims 14 to 17, wherein the time interval corresponds to the time difference between the time of occurrence of the reference network event and the time of occurrence of the transmitter application event.
19. A transmitter according to any one of claims 14 to 17, wherein the time interval corresponds to the time difference between the time of occurrence of the reference network event and the expected time of occurrence of the corresponding receiver application event, the expected time of occurrence of the corresponding receiver application event corresponding to the time of occurrence of the transmitter application event plus a given synchronization delay.
20. A receiver adapted for triggering an application event, the receiver being connected to a transmitter through a communication network, the knowledge of at least some of the next expected network events being shared with the transmitter and potentially with other receivers connected to the transmitter wherein the receiver comprises: -a receiving module for receiving synchronization information including a time interval and an index of a reference network event among the next expected network events; -a determining module for determining a waiting time to wait after the time of occurrence of the reference network event based on the received time interval; and -a triggering module for triggering the application event at the end of the waiting time after the occurrence of the reference event.
21. A receiver according to claim 20, wherein, the received time interval being the time intcrval between thc timc of occurrence of a transniittcr application cvcnt and the time of occurrcncc of thc rcfercncc network event, thc waiting time is determined by taking a given synchronization delay minus the received time interval.
22. A receiver according to claim 20, wherein, the received time interval being the time interval between the time of occurrence of the reference network event and the expected time of occurrence of the corresponding receiver application event which time of occunence corresponds to the time of occurrence of the transmittcr application event plus a given synchronization delay, the waiting time is determined by taking the value of the received time interval.
23. A recciver according to any onc of claims 20 to 22, whcrein thc rccciver furthcr 1 5 comprises: -an estimating module for cstimating the timc of occurrence of the rcference network event, the network access being controlled by a CSMA/CA scheme, based on a shared knowledge of the backoff values used by the transmitter and the rcceivers participating in thc method.
24. A receiver according to claim 23, wherein the receiver further comprises: -a determining module for determining the difference between the actual time of occunence of the reference network event and the estimated time of occurrcncc of the rcfcrence network evcnt; and -an adjusting module for adjusting the waiting time accordingly.
25. A receiver according to claim 23 or 24, wherein the estimating module is adaptcd for estimating anew the time of occurrcncc of the reference network event whcn a node gcncrates an uncxpcctcd network event bctwccn the previous estimation and the actual time of occurrence of the reference network event; and the receiver further comprises: -an updating module for updating the waiting time accordingly.
26. A transmission system for synchronizing an application event between a transmitter connccted to at least one receiver through a communication network, wherein the transmission system comprises: -a transmitter for sending synchronization information relating to the application event by thc transmitter according to any one of claims 14 to 19; and -at least a receiver for triggering an application event, based on said synchronization information, according to any one of claims 20 to 25.
27. A computer program product for a programmable apparatus, the computer program product comprising a sequence of instructions for implementing a method according to any one of claims 1 to 13, when loaded into and executed by the programmable apparatus.
28. A computer-readable storage medium storing instructions of a computer program for implementing a method according to any one of claims Ito 13.
29. A method of synchronisation substantially as hereinbefore described with reference to, and as shown in Figures 3 to 7.