WO2026012829A1 - Fairness restoration for txop pre-emption - Google Patents

Abstract

To restore fairness compared to legacy stations, the STA resets the current value of the backoff counter of the concerned AC (of the data transmitted during the pre-empted period) before starting using again the backoff counter to perform channel access. The reset includes randomly selecting an initialization value for the backoff counter from [0, CW] where the current contention window CW is kept unchanged. The reset may be upon obtaining the TXOP pre-emption or upon successfully transmitting the data within the pre-empted period.

Description

FAIRNESS RESTORATION FOR TXOP PRE-EMPTION

FIELD OF THE INVENTION

The present disclosure relates generally to wireless communication and more particularly to medium access.

BACKGROUND OF THE INVENTION

The approaches described in this section could be pursued, but are not necessarily approaches that have been previously conceived or pursued. Therefore, unless otherwise indicated herein, the approaches described in this section are not prior art to the claims in this application and are not admitted to be prior art by inclusion in this section. Furthermore, all embodiments are not necessarily intended to solve all or even any of the problems brought forward in this section.

Wireless communication networks are widely deployed to provide various communication services such as voice, video, packet data, messaging, broadcast, etc. These wireless networks may be multiple-access networks capable of supporting multiple users by sharing the available network resources. Examples of such multiple-access networks include Code Division Multiple Access (CDMA) networks, Time Division Multiple Access (TDMA) networks, Frequency Division Multiple Access (FDMA) networks, Orthogonal FDMA (OFDMA) networks, and Single-Carrier FDMA (SC-FDMA) networks.

The WLAN (Wireless Local Area Network) technology based on the IEEE (Institute of Electrical and Electronics Engineers - RTM) 802.11 family standards provides a very simple distributed channel access mechanism. Distributed channel access means that a wireless device, in IEEE 802.11 terminology known as a station (STA), either access point (AP) or a non-access point (non-AP), tries to access an operating channel when it has data to send, usually using contention schemes on a so-called primary channel.

An operating channel of 40, 80, 160 or 320MHz bandwidth (as defined in the latest IEEE 802.11 be D6.0 standard, but may be wider in future amendments) is usually made of a primary channel and one or more secondary channels (each channel being 20MHz or a multiple thereof). The primary channel is used for signalling (including channel access procedure) and backwards compatibility while the secondary channels are only used when sending data at full speed.

The basic unit of allocation of the right to transmit onto the wireless medium is the transmission opportunity or “TXOP”, defined by a starting time and a defined maximum length.

The TXOP is often obtained by a STA winning an instance of contention. A contention function is in charge of driving the contention.

Known contention includes the Distributed coordination function (DCF). DCF relies on a carrier-sense multiple access with collision avoidance (CSMA/CA) with a binary exponential backoff algorithm. It uses a backoff counter which is initialized with a backoff value randomly drawn from respective contention parameters. The backoff counter is decremented during a contention period at each time slot the wireless medium is detected as idle. Conventionally, the contention period starts a DIFS (DCF InterFrame Space - equal to SIFS + 2 * SlotTime) after the medium is detected as being idle. The decrementing is stopped and deferred when the wireless medium becomes busy. On the other hand, when it reaches zero, the station gains access to the wireless medium, hence can transmit pending data.

Below, a “TXOP holder” or “TXOP owner” is understood as being a station (STA) that has either been granted a TXOP by the hybrid coordinator (HC) or successfully contended for a TXOP.

QoS (Quality of Service) is provided in 802.1 1 networks thanks to Enhanced Distributed Channel Access or "EDCA" which defines traffic categories and four corresponding access categories making it possible to handle differently high-priority traffic compared to low-priority traffic. Implementation of EDCA in the stations can be made using a plurality of traffic queues (known as "Access Categories (AC)") for serving data traffic at respective different priorities, each traffic queue being associated with a respective queue backoff counter.

EDCA enhances DCF. Conventionally, EDCA proposes four ACs, each one having its own queue backoff counter, the initializing backoff value of which being randomly drawn from respective queue contention parameters, known as EDCA parameters. The function performing the EDCA contention for one AC is also known as EDCA function or “EDCAF”. As the EDCA parameters are specific to each AC queue, packets from different ACs are transmitted according to different priorities mirroring the respective EDCA parameters.

Once the STA is granted a TXOP, the medium is reserved for that STA for the TXOP length, while the other STAs not involved in the TXOP cannot transmit. Indeed, during the TXOP, the granted STA organizes data transmissions, be them single-user transmissions, multi-user transmissions, uplink transmissions, downlink transmissions, peer-to-peer transmissions, and so on.

As all the STAs obtain access to the medium using the same channel access mechanism (DCF or EDCA), fairness between the STAs is ensured.

In a WLAN network in which multiple STAs are active, some STAs may not be granted opportunities to transmit during successive TXOPs, hence delaying their traffic transmissions. This obviously degrades network performance, in particular with respect to low latency traffic such as high-definition video, advanced telemedicine, ultra-low latency gaming, and ARA/R (augmented/virtual reality) data.

Accordingly, improvements in the field are desired. That is why discussions have emerged in the IEEE802.11 bn - Ultra High Reliability (UHR) - work group to enable pre-emption of the wireless medium while a TXOP is on-going. Pre-emption would allow the medium to be accessed by interrupting a rightful transmission sequence (TXOP) by a contention function (e.g. EDCAF) that did not obtain the TXOP. A use case regards a higher-priority stream that could preempt a lower-priority stream during transmission in order to meet strict timing requirements.

This can potentially cause even worse performance degradation or at least fairness issues for traffic for which low latency is less important, at least in some instances.

Accordingly, further improvements in the field are desired.

SUMMARY OF THE INVENTION

Existing mechanisms to restore fairness already exist in the IEEE 802.11 family of standards.

This is the case of the MU-EDCA scheme where STAs gaining opportunities (resource units) during MU (multi-user) transmissions triggered by the AP have to modify their contention parameters (known as EDCA parameters) in order to favour Uplink (UL) MU operations over UL Single-User (SU) operations. In this perspective, the legacy EDCA parameters are replaced, at least temporarily, by so-called MU-EDCA parameters that are more restrictive than the legacy EDCA parameters to favour MU operations.

However, this fairness restoration scheme does not appear appropriate to TXOP preemption. Firstly, the MU-EDCA scheme substantially degrades the contention parameters of the STA which degradation is supposedly balanced by a fair behaviour from the AP to schedule the STA in future MU transmissions. In particular, the STA may be forbidden to use EDCA during a timeout period. However, the TXOP pre-emption is conducted by the STA itself, without counterbalancing by the AP. As a result, the network access for the STA would be highly degraded, contrary to the sought fairness and the requirements of its pre-empting traffic (which may be low latency traffic).

Secondly, the MU-EDCA parameters are defined for the entire BSS. Therefore, they are unlikely to be adapted to the specificities of each STA (of their pre-empting traffic).

Finally, the MU-EDCA scheme is driven by the AP whereas the TXOP pre-emption scheme can be obtain by the STA at any time. Similarly, the MU-EDCA scheme is only designed to provide opportunities to the STA to transmit UL data to the AP, whereas the TXOP pre-emption scheme should remain open to other types of data, e.g. to peer-to-peer (P2P) data.

In this context, the present disclosure proposes a new fairness restoration scheme to be applied when stations transmit traffic through TXOP pre-emption.

A first aspect of the disclosure relates to a communication method in a wireless network, comprising, at a station (STA), be it an AP or a non-AP STA within a multi-link device (MLD) or not: pre-empting a TXOP gained by a TXOP holder, for the STA to transmit data during a preempted period. TXOP pre-empting means pre-empting medium access over the TXOP holder in the TXOP; and applying a modifying operation on a current value of a backoff counter after having preempted a TXOP, before starting using the backoff counter to perform channel access. The backoff counter may be usually used by the STA to perform channel access and obtain a transmission opportunity (TXOP) for itself.

By modifying the BC (backoff) value, and not necessarily the EDCA parameters, the described scheme proposes to compensate each single TXOP pre-emption obtained by the STA as it was a new SU (Single User) transmission opportunity obtained by the STA. It results that the proposed fairness restoration scheme is fair with regards to the existing mechanisms, meaning it does not introduce less or more transmission opportunities for the STA compared to the other (legacy or not) STAs of the BSS.

It may be noted, as further described below, that the proposed scheme can be applied to multiple backoff counters BCs (e.g. those of the four EDCA ACs) on a BC basis as long as data of the BC are transmitted in the TXOP pre-emption.

Correlatively, the invention also provides a wireless communication device comprising at least one microprocessor configured for carrying out the method above.

Optional features are defined below with reference to methods, while they can be transposed into device features.

In some embodiments, the modifying operation includes drawing a new initializing value for the backoff counter. This approach ensures the proposed fairness restoration scheme provides the same fairness as the DCF or EDCA mechanism.

In particular embodiments, in case the new initializing value is lower than the current value of the backoff counter, the current value of the backoff counter is kept unchanged to perform the next channel access. This is to avoid having the STA to take benefits of the proposed scheme to obtain earlier new SU transmission opportunities.

In other particular embodiments, the new initializing value is drawn from an unmodified contention window. It means the current value CW of the contention window is not modified I changed by the TXOP pre-emption. The STA therefore keeps its legacy behaviour when EDCA contending after the TXOP pre-emption. This is to maintain equal fairness with the other STAs. It is also because while the CW is usually updated according to EDCA contention success or failure to reflect the EDCA medium access conditions, TXOP pre-emption is substantially a different access scheme not impacted by the EDCA medium access conditions.

In other embodiments, the modifying operation includes multiplying the current value of the backoff counter by a factor strictly higher than 1 , e.g. an integer factor.

In some embodiments, applying the modifying operation on the current value of the backoff counter is in response to successfully pre-empting the TXOP. The proposed scheme therefore is perfectly aligned with the conventional behaviour of a STA gaining a SU communication access.

In alternative embodiments, applying the modifying operation on the current value of the backoff counter is in response to successfully transmitting data within the pre-empted period. Fairness is therefore restored only in case the STA takes benefits of the TXOP pre-emption. In particular, the applying of the modifying operation may be further dependent on whether there is still data in the traffic queue corresponding to the backoff counter or not. Indeed, there is no need to use a new backoff value in case there is no longer data to be transmitted in the concerned AC. On the contrary, the backoff counter may be disabled, up to when new data to be transmitted arrive.

In some embodiments, the STA includes a main traffic queue and a separate alternate traffic queue both associated with the backoff counter, the alternative traffic queue being used to store pre-empting data to be transmitted during pre-empted periods while the main traffic queue is used to store legacy data to be transmitted in TXOPs obtained through channel access, wherein the modifying operation is applied depending on whether there is still data in the alternate traffic queue corresponding to the backoff counter or not. This approach seeks to avoid degrade the pacing for the AC queue in case only legacy data remain.

In some embodiments, the TXOP holder is a station different from the STA.

In variants, the STA is the TXOP holder that gained the TXOP for a given type of data and pre-empts its own TXOP to transmit another type of data. This defines a self TXOP preemption.

In some embodiments, the method further comprises, at the STA, transmitting to an access point (AP) to which it is associated, an unsolicited Buffer Status Report (BSR) frame to report a remaining queue size for the traffic queue corresponding to the data transmitted during the pre-empted period. This approach provides better network efficiency, since the AP can, thanks to the reported BSR, adjust dynamically the UL resources it can schedule for its associated STAs.

Another aspect of the disclosure relates to a communication method in a wireless network, comprising, at a station (STA): using a first backoff counter to perform channel access and obtain a transmission opportunity (TXOP) to send first data associated with the first backoff counter; in case there is no more first data to be sent while the TXOP is on-going, sending second data associated with a second backoff counter, within the TXOP; and applying a modifying operation on a current value of the second backoff counter after having transmitted the second data and before starting using (again) the second backoff counter to perform channel access.

This configuration allows fairness to be restored in case a TXOP holder can transmit data from an AC different from the one used to contend for channel access and obtain the TXOP.

Another aspect of the disclosure relates to a non-transitory computer-readable medium storing a program which, when executed by a microprocessor or computer system in a wireless device, causes the wireless device to perform any method as defined above.

At least parts of the methods according to the disclosure may be computer implemented. Accordingly, it may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident 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". Furthermore, it may take the form of a computer program product embodied in any tangible medium of expression having computer usable program code embodied in the medium.

Since the proposed mechanisms can be implemented in software, they can be embodied as computer readable code for provision to a programmable apparatus on any suitable carrier medium. A tangible, non-transitory 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 device and the like. A transient carrier medium may include a signal such as an electrical signal, an electronic signal, an optical signal, an acoustic signal, a magnetic signal or an electromagnetic signal, e.g., a microwave or RF signal.

BRIEF DESCRIPTION OF THE DRAWINGS

Some embodiments of the present disclosure are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings, in which like reference numerals refer to similar elements and in which:

Figure 1 illustrates a typical wireless communication system in which embodiments of the disclosure may be implemented;

Figure 2 illustrates 802.1 1 e mechanism for the backoff counter countdown in a conventional channel access scheme;

Figure 3 illustrates, using a flowchart, general steps of a communication method involving a fairness restoration to compensate TXOP pre-emption, according to embodiments;

Figure 4 illustrates, using a timeline, an exemplary scenario of TXOP pre-emption according to embodiments;

Figure 4a illustrates, using a timeline, another exemplary scenario of TXOP pre-emption according to embodiments;

Figure 4b illustrates, using a timeline, another exemplary scenario of TXOP pre-emption according to embodiments;

Figure 5 illustrates, using a flowchart, steps of a communication method involving a fairness restoration to compensate TXOP pre-emption, according to embodiments;

Figure 6a shows a schematic representation of a communication device; and

Figure 6b illustrates schematically the architecture of the communication device of Figure 6a.

DETAILED DESCRIPTION

The invention will now be described by means of specific non-limiting exemplary embodiments and by reference to the figures.

In the following description, the term legacy refers devices that may operate in accordance with one or more of IEEE 802.11 a/b/g/n/ac/ad/af/ah/aj/ay/ax/be, or another legacy wireless communication standard. The legacy devices may be STAs, IEEE STAs or Wireless- Fidelity (Wi-Fi) STAs.

An AP may communicate with legacy devices in accordance with legacy IEEE 802.11 communication techniques.

Figure 1 illustrates a communication system in which several communication devices or stations (or “nodes”) 101-107 exchange data frames over a radio transmission channel 100 of a wireless local area network (WLAN), under the management of a central station, or access point (AP) 1 10, also seen as a station of the network. The radio transmission channel 100 is defined by an operating frequency band constituted by a single channel or a plurality of channels (usually each of 20MHz width) forming a composite or operating channel. Below the “medium” or “wireless medium” is considered synonymous with the radio transmission or composite or operating channel.

In the following, the word “station” or “STA” refers to any kind of station. The wording “access point station”, or in short “access point” (AP), refers to the station playing the role of access point 110. The wording “non-access point station”, or in short “non-AP station” or “non-AP STA”, refers to the other stations 101 -107.

The STAs may be affiliated stations of multi-link devices as defined in the IEEE P802.11 be/6.0 standard, and/or stations implementing multi-user transmission features as defined in the IEEE 802.11 ax standard, and/or stations implementing any previous version of the 802.11 standard.

Access to the shared radio medium to send data frames is primarily based on the CSMA/CA technique, for sensing the carrier and avoiding collision by separating concurrent transmissions in space and time.

Carrier sensing in CSMA/CA is performed by both physical and virtual mechanisms. Virtual carrier sensing is achieved by transmitting control frames to reserve the medium prior to transmission of data frames.

Next, a source or transmitting station, including the AP, first attempts through the physical mechanism, to sense a medium that has been idle for at least one interframe time period known as DIFS (standing for DCF InterFrame Spacing), before transmitting data frames.

However, if it is sensed that the shared radio medium is busy during the interframe period, the source station continues to wait until the radio medium becomes idle.

The wireless communication system of Figure 1 comprises physical access point 110 configured to manage the WLAN BSS (Basic Service Set), i.e., a group of non-AP STAs which have previously registered to the AP. Such BSS managed by the AP is called an infrastructure BSS. In the following, the term BSS will be used as an equivalent of infrastructure BSS.

Once the BSS is established, the Access Point can bridge traffic inside the BSS or from other networks (e.g., wired networks) into the BSS (or vice and versa). Thus, the non-AP STAs of the BSS originally talked to the AP only, which is in charge of relaying data frames if the data frames are targeted to another non-AP STA of the BSS. Two directions of communication are therefore defined: “downlink” from the AP to the non-AP STAs and “uplink” from any non-AP STA to the AP.

In embodiments, the AP may consider setting up a pre-emption enabled BSSID of a multiple BSSID set, where the TXOP pre-emption scheme as proposed below can be effective. Preferably, this BSS is configured with long TXOP Limits. In theory, this would affect the medium access delay for legacy stations, in contrary to TXOP pre-emption enabled stations of the disclosure that can pre-empt those TXOPs.

Recent developments in the 802.1 1 family of standards have given the opportunity to the non-AP STAs to send data directly to another non-AP STA, referred to as peer-to-peer (P2P) or Direct Link communications.

The basic medium access to obtain a transmission opportunity or “TXOP” is the DCF channel access scheme that uses CSMA/CA and a random backoff count following a busy medium condition, i.e., an end of a previous frame transmission (by any STA).

The time interval between frames is called the interframe space (IFS). A STA determines that the medium is idle through the use of the CS (Channel Sensing) function for the interframe space specified. The interframe space corresponds from the end of the last symbol of the previous frame to the beginning of the first symbol of the preamble of the subsequent frame as sensed on the wireless medium.

Figure 2 illustrates some interframe spaces to provide priority levels for access to the wireless medium.

Not shown, the reduced interframe space or RIFS was originally used to reduce overhead and thereby increase network efficiency, in replacement of the SIFS to separate multiple transmissions from a single transmitter, when no SIFS-separated response transmission is expected. RIFS is the shortest IFS, e.g. 2 ps.

The short interframe space or SIFS is used prior to transmission of specific frames, such as Ack (acknowledgement) frames, listed in the standard IEEE Std 802.11 ™-2020, section 10.3.2.3.3. SIFS is for example 16 ps.

SIFS is used when STAs have seized the medium and need to keep it for the duration of a frame exchange sequence to be performed. Using the smallest gap (RIFS has become obsolete in the meantime) between transmissions within the frame exchange sequence prevents other STAs, which are required to wait for the medium to be idle for a longer gap, from attempting to use the medium, thus giving priority to completion of the frame exchange sequence in progress.

The priority interframe space or PIFS is used to gain priority access to the medium to transmit specific frames as listed in the standard IEEE Std 802.11 ™-2020, section 10.3.2.3.4. In particular, priority access is often offered to the AP and/or to a TXOP holder (a STA that is granted a TXOP). PIFS is for example 25 ps.

The DCF interframe space or DIFS is used by STAs operating under the DCF to transmit Data frames (MPDUs) and Management frames (MMPDUs). For example, a STA using may transmit if both after it has correctly received a frame, its Carrier Sensing (CS) mechanism determines that the medium still idle at the end of the DIFS period and the STA’s backoff counter BC has a value of zero. DIFS is for example 34 ps.

A STA invokes the backoff procedure to transmit a frame.

After the DIFS medium idle time, a contention period starts. The STA generates a random backoff count for an additional deferral time before transmitting, unless the backoff counter already contains a nonzero value. The initializing backoff count value is a pseudorandom integer drawn from a uniform distribution over the interval [0, CW], where CW (contention window) is an integerwithin the range [CWmin, CWmax], As a shortcut, [0, CW] is also referred to as “contention window”.

The backoff procedure uses the CS mechanism to determine whether there is activity on the medium during each backoff slot “aSlotTime” (9ps). If no medium activity is indicated for the duration of a particular backoff slot, then the backoff procedure decrements its backoff counter. If the medium is determined to be busy, the backoff counter is not decremented for that slot, and the backoff counter can be next decremented only during a next contention period, i.e., after the medium has been determined to be idle for the duration of a DIFS. Transmission commences when the backoff counter equals 0.

Management of quality of service (QoS) has been introduced at station level in the wireless networks, through well-known EDCA mechanism defined in the IEEE 802.1 1 e standard (Enhanced Distributed Channel Access). EDCA enhances or extends the functionality of the original DCF scheme.

EDCA adds four independent enhanced distributed channel access functions (EDCAFs) to provide differentiated priorities to transmitted traffic, through the use of four different access categories (ACs), each having its own transmit queue. The four ACs are the following in decreasing priority order: voice (or “AC_VO”), video (or “AC_VI”), best effort (or “AC_BE”) and background (or“AC_BK”). In embodiments, QoS may be extended to support a different - higher - number of access categories.

A mapping of UP (user priorities of MSDUs, incoming from an upper layer) to the transmit queue and the mapping to AC is well-known and not reproduced here for brevity.

A backoff procedure (as described above for the DCF scheme) can be invoked by each EDCAF (i.e., independently for each AC).

With EDCA, the DIFS period is replaced at AC level by a so-called arbitration interframe space or AIFS. An AIFS is thus defined for each AC: AIFS[AC],

A STA using the EDCAF obtains a TXOP for an AC if the STA’s CS mechanism determines that the medium is idle during the AIFS[AC] period, after a correctly received frame, and the backoff counter for that AC, BC[AC], has a value of zero. The duration AIFS[AC] is a duration derived from value AIFSN[AC] (which is between 2 and 15): AIFS[AC] = AIFSN[AC] * aSlotTime + SIFS.

Each EDCAF[AC] maintains its backoff counter BC[AC], which has a value measured in backoff slots. After the AIFS[AC] period, the backoff counter is set to an integer value chosen randomly with a uniform distribution taking values in the range [0, CW[AC]], unless the backoff counter already contains a nonzero value. Contention window for the AC, CW[AC], is an integer within the range [CWmin[AC], CWmax[AC]]. As for DCF, if no medium activity is indicated for the duration of a particular backoff slot (aSlotTime), then EDCAF[AC] decrements its backoff counter BC[AC], If the medium is determined to be busy, the backoff counter is not decremented for that slot, and the backoff counter can be next decremented only during a next contention period, i.e., after the medium has been determined to be idle for the duration of an AIFS[AC], Transmission commences when the backoff counter equals 0.

In a BSS, the EDCA Parameters, such as CWmin[AC], CWmax[AC], AIFSN[AC] and TXOP_Limit[AC] (maximum length of a TXOP a STA may request), are advertised by the AP in a so-called EDCA Parameter Set element in Beacon and Probe Response frames transmitted by the AP. The EDCA Parameter Set element therefore provides the EDCA Parameters for all (four) ACs.

Due to different EDCA Parameters for different ACs, the deferring of the contention period for the ACs are different. For example, AC_VO has the highest priority and as such has the lowest AIFS. Although configurable, default values of the AIFS are the following ones:

AC_VO 1 SIFS + 2 * slot time (AIFSN[3] = 2)

AC_VI 1 SIFS + 2 * slot time (AIFSN[2] = 2)

AC_BE 1 SIFS + 3 * slot time (AIFSN[1] = 3)

AC_BG 1 SIFS + 7 * slot time (AIFSN[0] = 7)

For example, as shown in Figure 2, two AIFS corresponding to AC=i and AC=j are considered. Due to this prioritizing difference, AC ‘j’ starts decrementing its backoff value earlier than less-prioritized AC T.

Relative prioritization of the ACs (hence QoS) is therefore obtained, which can be tuned by adjusting the EDCA Parameters of the ACs.

The ACs within the same STA thus compete one with each other (using their EDCAF) to access the wireless medium and to obtain a TXOP.

Furthermore, the use of lower AIFSN values, additional to the use of an on-average lower CW for high priority ACs compared to low priority ACs makes that traffic of a high priority AC has a higher chance to be transmitted than traffic from a low priority AC: a STA having high priority AC traffic statistically waits less, on average, to be granted a TXOP and then send its packet than a STA having low priority AC traffic.

Recently-created IEEE 802.11 bn (the successor of IEEE 802.11 be) Task group works on ultra-reliable low-latency communication (URLLC) requirements, focusing on some aspects such as tail latency and jitter, and high priority access for latency-sensitive applications.

Applications with data having tight latency requirements include high-definition video, advanced telemedicine, ultra-low latency gaming, and AR/VR (augmented/virtual reality) data.

Time-Sensitive Networking (TSN) is another exemplary targeted extension of 802.11 /WiFi communications with deterministic elements and time-sensitive capabilities. TSN enables reliable and timely communication over networks by providing a suite of mechanisms and protocols that enable networked devices to exchange critical data with very low latency and high reliability.

TXOP Pre-emption is proposed to reach tight latency requirements. TXOP Pre-emption is a mechanism to access the medium by interrupting a rightful transmission sequence by an EDCAF that did not obtain an on-going TXOP.

TXOP Pre-emption should operate with any type of communication and devices, including either or both of single-user (SU) or multi-user (MU) communication frames, either or both of uplink (UL) or downlink (DL) communication frames, either or both of single-link or multi-link devices.

As the STAs benefiting from TXOP Pre-emption have more opportunities to transmit compared to legacy STAs, fairness between the STAs is distorted.

Another situation raised fairness issues in the past, namely the MU feature of IEEE 802.11 ax. This feature also provides transmission opportunities (resource units in trigger-based MU communications) additional to the EDCA-based opportunities. When HE STAs are UL MU capable, they can either access the channel with their EDCAF to send an SU PPDU to the AP, or be scheduled in UL MU by the reception of a trigger frame from the AP.

Fairness is restored by forcing the HE STAs to modify their EDCA parameters, in order to favour UL MU operations in place of UL SU operations. Legacy STAs, however, should still be able to use normal EDCA parameters, in order not to be disfavoured. The first set (EDCA parameters) is referred to as legacy EDCA parameters that are for SU operations and may be used by both HE STAs and legacy STAs without MU capability. The second set of parameters is referred to as MU EDCA parameters which are defined to be more restrictive than legacy EDCA parameters in favouring MU operations.

When an HE STA receives a trigger frame scheduling the HE STA for an AC, the HE STA switches to MU EDCA parameters only for that AC. The HE STA then basically trusts that the AP will be able to schedule the HE STA efficiently for that particular AC traffic. The HE STA may cooperate by indicating regularly the evolution of its buffer status to the AP, e.g., by piggybacking this information with its data. In order to have a way out in case the AP is not efficient, the HE STA may switch back to the legacy EDCA parameters if the HE STA has not been scheduled after a pre-defined timeout period after the last time the HE STA was scheduled in UL MU.

Switching to the MU EDCA parameters in case of TXOP pre-emption does not sound appropriate for the following reasons.

Firstly, the MU-EDCA scheme substantially degrades the contention parameters of the STA which degradation is supposedly balanced by a fair behaviour from the AP to schedule the STA in future MU transmissions. In particular, the STA may be forbidden to use EDCA during a timeout period. However, the TXOP pre-emption is conducted by the STA itself, without counterbalancing by the AP. As a result, the network access for the STA would be highly degraded, contrary to the sought fairness and the requirements of its pre-empting traffic (which may be low latency traffic).

Secondly, the MU-EDCA parameters are defined for the entire BSS. Therefore, they are unlikely to be adapted to the specificities of each STA (of their pre-empting traffic).

Finally, the MU-EDCA scheme is driven by the AP whereas the TXOP pre-emption scheme can be obtained by the STA at any time. Similarly, the MU-EDCA scheme is only designed to provide opportunities to the STA to transmit UL data to the AP, whereas the TXOP pre-emption scheme should remain open to other types of data, e.g. to peer-to-peer (P2P) data.

Another fairness restoration scheme is proposed in the present disclosure, adapted to TXOP pre-emption. It involves updating the BC (DCF BC or any EDCA BC[AC] orthe like) in case of TXOP pre-emption to ensure better fairness while not claiming a further TXOP immediately. In particular, by only updating the BC without modifying the corresponding contention window CW (nor CWmin/CWmax), the proposed scheme does not modify the STA’s behaviour to ‘real’ medium accesses (outside the pre-empted TXOP).

Figure 3 illustrates, using a flowchart, general steps of a communication method involving a fairness restoration to compensate TXOP pre-emption, according to embodiments. The method may be implemented by any STA using the wireless medium currently reserved by a TXOP holder through an on-going TXOP. The STA may be a communication partner of the TXOP holder during the TXOP or be a STA not involved in the TXOP, or even be the TXOP holder in some embodiments. More generally, the STA may be a non-AP STA willing to send uplink data to the AP, a non-AP STA willing to transmit peer-to-peer data to a peer non-AP STA partner, an AP willing to send downlink data to one or more non-AP STAs, or even a second AP willing to obtain the medium for its BSS in a context of multi-AP coordination.

Step 300 mirrors conventional behaviour of a STA operating in the wireless medium, hence using its backoff counter or counters to perform channel access and obtain a transmission opportunity (TXOP). DCF uses one backoff counter that is decremented over time, while EDCA uses up to four backoff counters (one per EDCAF) that are decremented over time. The backoff counters are suspended when the medium becomes busy, e.g., when another STA gains a TXOP.

Step 310 occurs when one TXOP is on-going, gained by a TXOP holder.

The STA can then decide to pre-empt the on-going TXOP, i.e., the STA pre-empts medium access over the TXOP holder in the on-going TXOP. This is for the STA to transmit data during a pre-empted period within the TXOP.

A TXOP pre-emption means that the communications targeted by the TXOP are suspended during the TXOP, to the benefit of another communication not initially targeted by the TXOP. Therefore, the targeted communication can restart after the TXOP pre-emption, provided that enough time remains in the TXOP.

In the present disclosure, a communication targeted by the TXOP and a communication not targeted by the TXOP relate to traffic data managed by separate backoff counters, either because they are handled by separate stations or because they relate to separate ACs or the like.

Various embodiments to pre-empt medium access during the TXOP are described below.

Once the STA has pre-empted the medium - in replacement of the TXOP holder -, the STA may use the medium for its own transmissions - not initially targeted by the TXOP - as described below. The STA may initiate one or more frame exchange sequences.

Next, step 320 is applied to restore fairness of the STA compared to other STAs of the same wireless network (BSS). To do so, it applies a modifying operation on the current value of the backoff counter (in particular the one or ones having allowed the TXOP pre-emption, hence of the not initially targeted data) after having pre-empted a TXOP and before starting using again the backoff counter to perform channel access.

As a result, the STA will contend for new SU operations with a new backoff counter mirroring restored fairness. Indeed, as the pre-emption can be considered as an additional (but single one) SU operation, the backoff counter can merely be reinitialized, cancelling any beneficial decrement thereof that could have happened beforehand. A local fairness policy is implemented.

As apparent from the exemplary scenario below of TXOP pre-emption, TXOP preemption is particularly suitable for STAs that have latency sensitive traffic they wish to prioritize. Of course, it may be transposable to any type of traffic, whose transmission is to be prioritized over other types of traffic that are targeted by the on-going TXOP.

Figure 4 illustrates, using a timeline, an exemplary contention mechanism for TXOP preemption.

The Figure shows the behaviour of two non-AP STAs and one AP.

AP and STA1 are involved in an on-going TXOP 400 while STA2 is the station willing to pre-empt the TXOP over the TXOP holder (here the AP). Hence, STA2 implements the preemption mechanism. This is only an example. For instance, STA1 could alternatively (or simultaneously) implement the pre-emption mechanism to have control over the TXOP, in replacement of AP. Any other combination is possible, whateverthe TXOP holder is an AP or not.

For the sake of illustration, this timeline is illustrated for only one stream per STA, but of course several streams belonging to different priorities can coexist in a given STA. In that case, an optimization can be performed at the STA to arbitrate in between the streams and apply the TXOP pre-emption mechanism to the higher-priority stream.

The Figure illustrates an existing TXOP 400 in between AP and STA1 . TXOP 400 is defined as a preemptable TXOP, that is to say STAs may try to pre-empt it. AP and STA1 have a frame exchange sequence during which frame 401 is sent by AP to STA1 , in response to which STA1 sends an ACK frame 402 after a SIFS period.

Pre-empting STA2 senses the end of the ACK frame 402, ending the frame exchange sequence. It decides to start contending for pre-emption medium access by sensing the channel (or partial channel) during a pre-emption period 410 starting immediately afterthe end of the ACK frame 402 and lasting during the conventional interframe period. Hence, the pre-emption period 410 lasts at most the interframe period considered (a SIFS period in the present example).

As shown, a "pre-emption" backoff procedure is implemented to try to gain channel access. STA2, as pre-empting STA, only senses the channel condition during the pre-emption period 410. If the channel condition is idle for a pre-emption backoff slot time 411 , then the preemption backoff (PBO) counter is decremented by one. Otherwise, the PBO counter is not decremented, and the contention period 410 stops (another STA may have won the pre-emption contention) and the decrementing can be resumed for a next TXOP pre-emption operation within the same TXOP. STA2 wins the pre-emption and gains channel access when its PBO counter reaches zero with the channel still idle.

In embodiments, STA2 transmits a pre-emption request frame 403 (PR frame) over the medium upon successfully contending access to the medium within the interframe period. This request allows pre-empting STA2 to reserve a pre-empted period 420 within the on-going TXOP to perform its transmission. The pre-empted period 420 may end before the TXOP (as shown in the Figure) to give the medium back to the TXOP holder. In variants, STA2 may decide occupying the entire remaining time of the on-going TXOP.

Next, after the legal interframe space (SIFS period), STA2 can transmit another frame 404 (e.g. management frame, or data frame).

Optionally (as shown in dotted lines), the TXOP holder (here the AP) can acknowledge the pre-emption, so that STA2 waits for and receives, from the pre-empted TXOP holder, a preemption response frame 405 acknowledging the pre-emption request frame. This is to ensure the successful pre-emption. As shown, the ACK 405 is sent a SIFS after PR frame 403, and STA2 may start transmitting its data (frame 404) a SIFS after the ACK 405.

In other embodiments, STA2 transmits a data (or management) frame directly upon successfully contending access to the medium within the interframe period. In these embodiments, no PR frame 403 is sent.

In the example of the Figure, the pre-emption period 410 is started after the ACK 402 in order not to interrupt the transmission of an individual data frame 401 (which requires acknowledgment). In this example, an end of frame transmission (from which the contention starts) matches an end of an acknowledgement within a frame exchange sequence. However, in less strict implementations, the pre-emption period 410 may take place during the SIFS period immediately following data frame 401 .

Once the STA has performed the data exchange 404, it can release the pre-empted TXOP, meaning the TXOP holder can take it back to use it. The release takes place at the end of the pre-empted period 420. If the length of the pre-empted period 420 is signalled in the PR frame 403, the TXOP holder only has to wait for the end of the pre-empted period 420. If it is not signalled, the TXOP holder may consider using a PIFS period after each transmission within the pre-empted period 420 (to try to take the TXOP back) whilst the pre-empting STA uses only SIFS periods as long as it wants to continue transmitting in the pre-empted period 420. In summary, Figure 4 shows a scenario where the STA senses an end of frame transmission in a frame exchange sequence scheduled by a TXOP holder different from the STA within an on-going transmission opportunity (TXOP) granted to the TXOP holder. Next, the STA starts and performs for medium access during the interframe space or period within the on-going TXOP, immediately following the end of the frame transmission. The interframe period to be considered is the conventional one before a STA (e.g., the TXOP holder) involved in the TXOP sends the next frame. Usually, it is a SIFS period. However, in embodiments, a PIFS period may be contemplated. The contention may invoke a backoff procedure (PBO procedure as mentioned above). The success of the contention determines whether the STA has indeed pre-empted medium access over the TXOP holder in the TXOP or not. Once the STA has pre-empted the medium - in replacement of the TXOP holder -, the STA may use the medium for its own transmissions or to organize transmissions with other STAs.

Even if medium contention is a well-known technique, this contention-based TXOP preemption scheme through backoff procedure is original due to the execution of the contention within the on-going TXOP.

Compared to a legacy EDCA backoff procedure, the pre-emption backoff procedure is performed inside a TXOP, with shortest duration for backoff slots, in order that the maximum preemption contention period lasts an interframe delay (SIFS or PIFS). There is no need to wait for a deferring DIFS period before decrementing the pre-emption backoff counter because the contention is limited to small number of STAs (possibly under allowance by the AP via a QoS traffic management) and the medium was already granted to the TXOP holder which agrees to be pre-empted by those stations/flows.

One major advantage is that if no pre-emption is granted, the wireless medium is still kept by the TXOP owner and transmissions can continue in a legacy way.

As apparent from this scenario of Figure 4, the AC of the data in frame 404 have gained another opportunity to be sent, whilst the corresponding contention parameters are still alive and may result in a further medium access after the end of TXOP 400.

Figure 4a illustrates, using a timeline, another exemplary scenario forTXOP pre-emption, in particular self TXOP pre-emption. Indeed, in this scenario, the pre-empting STA is the TXOP holder that gained the on-going TXOP for a given type of data (e.g. an AC) and pre-empts its own TXOP to transmit another type of data (e.g. a more prioritized AC).

The Figure shows the behaviour of two STAs and one AP. Any other combination is possible, whatever the TXOP holder is an AP or not. This timeline is illustrated for two streams residing at the TXOP holder.

AP and STA1 are involved in the on-going TXOP 400 to transmit data traffic from ACi (meaning the data coming from AC of index T), in particular DL data. Indeed, every time a new MSDU arrives in an empty ACi, the STA (here the AP) can access the channel with its legacy EDCA parameters (for AC) to transmit this MSDU in SU mode. This is represented by EDCA contention 450 for "primary" ACi, prior to obtaining TXOP 400. TXOP 400 is defined as a preemptable TXOP, that is to say STAs may try to pre-empt it. AP and STA1 have a frame exchange sequence during which frame 401 is sent by AP to STA1 , in response to which STA1 sends an ACK frame 402 after a SIFS period.

The TXOP may have been reserved for that particular ACi, meaning the TXOP holder (AP) should not be allowed to transmit data traffic from another ACj (meaning the data coming from AC of index ‘j', j being different from i).

According to some embodiments, the TXOP pre-emption may be announced prior to or upon starting the on-going TXOP 400. As an example, the pre-empted period 420 may be announced through a Management frame by the AP, e.g., as a TWT period. Indeed, such TWT I pre-empted period 420 can be interpreted by beneficiary STAs as the period where the STAs have to wake up and look for a new frame emitted through pre-emption.

As an example, the timeline shows that the AP allows the interruption of its ongoing DL data 401-402 (AC) to transmit other DL data 414-415 from higher-priority (at least in term of transmission delay for the pending data) ACj targeted to same or other STAs, here STA2.

The timeline illustrates the reception (from upper layer) of ACj MSDU data during the TXOP 400. Of course, the data may have been received before TXOP 400. What matters is that the TXOP holder targets to send the data prior to an expiration date (delay bound), and so it may want to interrupt the legacy transmission of ACi data.

After the pre-empted period 420, the TXOP holder can get back its TXOP to continue sending targeted data, here frame 406 carrying ACi data.

As apparent from this scenario, ACj data have gained another opportunity to be sent, whilst the corresponding contention parameters are still alive and may result in a further medium access after the end of TXOP 400.

Figure 4b illustrates, using a timeline, another exemplary scenario for TXOP preemption, as described in document IEEE 802.11-23/1886.

In this scenario, low latency (LL) STAs send a Pre-emption Traffic Indication PRI 423 a SIFS after acknowledgment 402 of a DL transmission 401 from the AP. PRI could be a CTS frame.

Next, the LL STAs contend (410) to send UL LL data 424, which are acknowledged 425.

The PRI-contention-sending scheme can be repeated by the LL STAs, provided the PRI is sent a SIFS after a previous acknowledgment 402 or 425.

The AP can take back the medium a PIFS after any acknowledgment if no PRI 423 has been sent, to send any other DL frame 406. Compared to the timeline of Figure 4a, this timeline mandates that the regular interframe space becomes a PIFS (longer than legacy SIFS), which may appear detrimental to network efficiency.

As apparent from this scenario, the AC of the data in frame 404 have gained another opportunity to be sent, whilst the corresponding contention parameters are still alive and may result in a further medium access after the end of TXOP 400. As mentioned above, this additional opportunity can be unfair compared to other ACs. Furthermore, it can also be detrimental to network efficiency: indeed, the queue of the preempting data may become empty due to the additional opportunity, resulting in the next medium access obtained for this queue be wasted with no transmission. Therefore, there is a need not to degrade too much the legacy/primary-AC communications by still keeping the same medium access pace. A balance of the additional opportunity by modifying the legacy EDCA backoff counter for the concerned AC is proposed.

Figure 5 illustrates, using a flowchart, steps of a communication method involving a fairness restoration to compensate TXOP pre-emption, according to other embodiments. The method may be implemented by any STA wishing to use the wireless medium currently reserved through an on-going TXOP. The STA may be a communication partner of the TXOP holder during the TXOP or be a STA not involved in the TXOP, or be the TXOP holder itself as exemplified in Figure 4a for example. More generally, the STA may be a non-AP STA willing to send uplink data to the AP, a non-AP STA willing to transmit peer-to-peer data to a peer non-AP STA partner, or an AP willing to send downlink data to one or more non-AP STAs.

The method is presented for the transmission of a single data frame through pre-emption of an on-going TXOP. However, it is obvious that several frames pertaining to many classifications (ACs, SCSs as described below) may be considered when a TXOP pre-emption is gained.

The method starts at step 500 where the STA retrieves authorization information about which TXOP is open to pre-emption and/or which STA or STAs are allowed to pre-empt the ongoing TXOP.

In some embodiments, the pre-emption may be available for any TXOP and/or to the benefit of any STA, in which case step 500 can be omitted.

In other embodiments, limitations or restrictions on preemptable TXOP and pre-empting STA can be defined either statically (e.g., at the BSS level, the AP advertising its associated STAs using dedicated frames such as Beacon frames) or dynamically (e.g., by the TXOP holder itself).

Preferably, the TXOP holder signals whetherthe (on-going) TXOP is open to pre-emption (i.e., is available for pre-emption, meaning preemptable), e.g., in a management frame prior to the TXOP, in a frame reserving the TXOP or in a frame within the TXOP. For example, the AP, as TXOP holder, may allow an interruption of its ongoing (lower priority) DL data transmission (in the TXOP) by other STAs.

A management frame prior to the TXOP may include a TWT like frame announcing:

- the pre-emption contention period 410, hence indicating to the beneficiary (i.e., candidate to pre-emption) STAs to wake up, to check they have a frame to emit by pre-emption and then to perform pre-emption; or

- the pre-empted period 420, hence indicating to the beneficiary STAs to wake up and to check they receive a frame during the pre-empted period.

A frame reserving the TXOP may include a Trigger frame or a RTS (or MU-RTS) frame. A frame within the TXOP may include the first frame (MPDU) sent by the TXOP holder within the TXOP, regardless of the type of frame. Alternatively, it may be any frame of a frame exchange sequence after which the pre-emption is allowed.

A MAC signalling may be used, providing the pre-emption authorization or prohibition in one or more reserved bits of the MAC header, in a new A-Control field, or in one or more reserved bits of a MPDU delimiter. It is possible for the MAC signalling (indicating pre-emption authorization or prohibition) to be inserted by the PHY layer.

Similarly, the TXOP holder may signal which STAs are allowed to pre-empt its TXOP, in the same type of frames (prior to the TXOP, the frame reserving the TXOP or a frame within the TXOP). Or the AP may signal the authorized STAs for TXOP pre-emption for the entire BSS, which authorization may be provided for all types of TXOP or per TXOP type (e.g., depending on the type of frame reserving the TXOP).

A similar MAC signalling may be used.

As an example, the authorization for TXOP pre-emption may be given to: all non-AP STAs associated with the TXOP holder acting as AP, or one or more STAs specifically identified by the TXOP holder, or the AP with which the TXOP holder acting as a non-AP STA is associated.

The definition of the authorized pre-empting STAs may also be given through an indication of the traffic authorized in the pre-empted period 420, which traffic has its own backoff counter. In other words, the TXOP holder (or alternatively the AP) may signal which type or types of pre-empting traffic are authorized for transmission in case of pre-emption of the TXOP. In that case, this is a duty of the candidate pre-empting STA to determine, based on the traffic they wish to transmit, whether they are authorised or not to perform the TXOP pre-emption. Non-limitative exemplary allowed pre-empting traffics, may include one or more User Priorities (UP), one or more Traffic classifications (TCLAS); one or more transmission directions (uplink traffic, downlink traffic, peer-to-peer traffic); one or more Stream Classification Service (SCS) streams; one or more specific DSCP mapping policies.

Next step is step 510 where the STA determines that a MAC layer frame requires transmission, hence TXOP pre-emption should the medium not be available (TXOPs are ongoing), for instance to meet latency requirements. This frame and any other data that can be transmitted are referred below as pre-empting data.

In one or more embodiments, time-sensitive (low latency) packets may be indicated by higher layers of the communication stack (e.g., based on user priority), and corresponding MSDU's may be identified to trigger the TXOP pre-emption scheme.

The STA may for example include a MSDU classification module configured to classify any MSDU received from higher levels of the communication stack as time-sensitive and thus placed it in a dedicated time-sensitive queue. It may be one of the four legacy EDCA queues or an additional one dedicated to pre-empting traffic which may share the EDCA parameters with one of the four EDCA queue. In the last case, only the additional time-sensitive queue is allowed to use the TXOP pre-emption scheme in case of emergency.

As mentioned above, some types of traffic data may be allowed for TXOP pre-emption and other types not allowed. The STA may have built, based on the information retrieved at step 500, a black list of ACs (or traffic types) forwhich TXOP pre-emption is not allowed and/or a white list of ACs for which TXOP pre-emption is authorized. For example, if the AC of pending data belongs to a predefined group of ACs, the TXOP pre-emption scheme may be disabled.

Similarly, the proposed TXOP pre-emption mechanism may be enabled/activated or disabled/deactivated on demand, by the AP possibly upon request of a non-AP STA. Any management frame may be used although the SCS Action frame sounds suitable for such signalling. As an example, the AP or the STA may transmit (to the other) a Stream Classification Service (SCS) Action frame including an SCS Descriptor element defining a class of data, the SCS Descriptor element having a field or bit set to a first value to enable the TXOP pre-emption backoff procedure for the class of data or set to a second value to disable such TXOP pre-emption for the class of data.

At step 520, the STA determines whether the TXOP pre-emption scheme is to be triggered. This may consist in determining whether a transmission opportunity (TXOP) granted to a TXOP holder is on-going (identified or sensed through the reception of a frame, such as an RTS-CTS exchange or a trigger frame, reserving the TXOP to the TXOP holder), then checking that the on-going TXOP is preemptable for the STA (authorized) and the pre-empting data to be transmitted (authorized), based on the information retrieved at step 500 and the data frame identified at step 510.

In some embodiments, additional conditions may be considered to trigger the TXOP preemption scheme for efficiency purposes.

For example, the STA may not systematically trigger the scheme but considers whether the remaining time in the on-going TXOP (thanks to the signalled duration of the TXOP) is reasonable or sufficient to transmit the pre-empting data (identified at step 510). To illustrate this, in case the authorization to pre-empt the TXOP is conveyed in a MPDU transmitted near the end of the TXOP, the pre-emption scheme may be triggered only if there is enough room (time) to transmit the pre-empting data (here an MPDU conveying the new MSDU data). This is to ensure that the pre-empting data can be transmitted within the duration of the current TXOP. As a result of test 520, the decision about triggering the TXOP pre-emption scheme may be based on a successful determination of whether the pre-empting data can be transmitted in priority during the TXOP.

Other additional criteria may be taken into account in combination or in variants, such as an occupancy threshold for the dedicated time-sensitive queue (or the like), a flow priority, and so on.

Another criterion relies on the number of TXOP pre-emption operations already performed in the considered on-going TXOP. In embodiments, it may be considered that the STA can contend for medium access multiple times (i.e., during multiple interframe periods) within the same TXOP up to a predefined maximum numberof contending tries. Afterthis maximum number of tries (named "PBO Retry Limit"), the STA is no longer authorized to perform TXOP pre-emption in this TXOP.

Yet another criterion relies on the existence of a previous successful TXOP pre-emption for the STA and/or for the type of pre-empting data, within the on-going TXOP. Indeed, in embodiments, a single successful pre-emption procedure may be allowed inside a TXOP per a given traffic flow and/or STA. This is because a pre-empting STA has still the opportunity to aggregate several MSDUs from different traffic flows inside the frame it will send when preempting the TXOP. This approach reduces the number of contentions and thus of possible collisions.

In case of no TXOP pre-emption is needed, the process keeps the current value of the BC unchanged to perform legacy EDCA (step 590) after which the process ends.

In case of TXOP pre-emption is needed, next step is step 530 where the STA performs TXOP pre-emption.

In case of self TXOP pre-emption (e.g., scenario of Figure 4a), the STA may only wait for a TWT period defining the pre-empted period 420.

In case of pre-emption of a TXOP owned by another STA (e.g., scenario of Figure 4a), the STA may sense an end of frame transmission within the on-going TXOP. This may be done through Channel Sensing (CS) of the medium by the STA. Upon detecting this end, the STA starts, at step 530, the TXOP pre-emption operation, in particular a pre-emption backoff procedure, within the legacy interframe period that follows the previous frame, e.g., a SIFS period following ACK 402 in the scenario of Figure 4. This step represents the contention phase of the TXOP pre-emption scheme. The PBO backoff procedure is performed by the queue backoff engine associated with the AC corresponding to the pre-empting data to be transmitted (as identified at step 510). The PBO counter is set to an initializing value for decrementing. The STA starts decrementing the PBO counter each elementary time unit (e.g. RIFS) the communication channel is detected as idle. If the medium is determined to be busy (meaning a concurrent preempting STA has issued a medium access for pre-emption, as described at step 540), the PBO counter is not decremented for that slot, the PBO counter is suspended and the PBO counter can be next decremented only during a next pre-emption contention period 410, preferably in the same on-going TXOP. The decrementing lasts as long as the interframe period 410 does not end and the medium remains idle and the PBO counter (or none if multiple PBO counters) does not reach zero. Pre-emption success is detected when the PBO counter (or one if multiple PBO counters for multiple ACs) reaches zero.

In case of pre-emption of a TXOP owned by the AP (e.g. scenario of Figure 4b), the STA may send a PRI frame 423 and then contend to send its pre-empting data.

In case of no successful TXOP pre-emption, the process keeps the current value of the BC unchanged to perform legacy EDCA (step 590) after which the process ends. In case of successful TXOP pre-emption, the STA can access the medium and use it at step 540. The TXOP pre-emption suspends the on-going communications in the TXOP, i.e., the transmissions initially targeted by the TXOP.

In embodiments as illustrated for example in Figure 4, the STA announces the TXOP pre-emption by sending a pre-emption request frame, PR frame 403, as soon as the PBO counter reaches zero. As the TXOP pre-emption is contention-based, frame collision may occur (if several pre-empting STAs are using the same pre-emption slots). The PR frame, together with the ACK 405 below, allow PR frame collision to be detected at low cost compared to a long data frame.

An exemplary PR frame may be a PPDU made of legacy fields used for 802.11 frame detection, synchronization, carrying necessary information (e.g., MCSs and frame length), such as the L-STF field, the L-LTF field and the L-SIG field. Using legacy fields ensures backward compatibility and short frames and easy detection by 802.11 stations.

Once the PR frame 403 has been transmitted, the pre-empting STA can initiate (and perform) one or more frame exchange sequences 404 (using pending pre-empting data) with any other station. Note that in case the pre-empting STA is the AP, it may schedule MU transmission (uplink or downlink).

Optionally, the STA can wait for a pre-emption response frame, ACK 405 in Figure 4, from the TXOP holder that validates the TXOP pre-emption. This is to reduce risks of collision during the data transmission 404. The pre-emption response frame may be a mere acknowledgment frame, i.e., Ack frame. In variant, the pre-emption response frame may be a copy of PR frame 403.

In other embodiments depicted in Figure 4a for example, the pre-empting STA may directly perform data transmission 414 upon gaining the pre-empted period 420.

Next to step 540, the STA applies a modifying operation on the current value of the (namely queue or EDCA) backoff counter, before starting using it again to perform channel access.

The Figure suggests that the modifying operation resets the backoff counter in a conventional way, i.e., it includes drawing a new initializing value for the backoff counter from [0, CW], Preferably, CW is unmodified as well as the EDCA parameters of the target AC (the AC for which TXOP pre-emption was successful) and the optional MU-EDCA parameters of the target AC.

For example, a new backoff computation is performed by the queue backoff engine associated with the target AC queue. Possibly a new backoff value is recomputed for several AC queues, e.g., all those that have been transmitted during the pre-empted period 420.

To avoid any benefits for the STA recomputing its backoff value, in case the new initializing value is lower than the current value of the backoff counter, the current value of the backoff counter is kept unchanged to perform the next channel access. This ensures next EDCA medium access will not be moved earlier. In this scenario, the modifying operation is applied on the current value of the backoff counter in response to successfully pre-empting the TXOP, in particular whatever the data transmission 540 is successful or not.

In variants however, the modifying operation is applied on the current value of the backoff counter in response to successfully transmitting data 404, 414 within the pre-empted period 420. This is illustrated in Figure 5 by the dashed arrow from step 540 where, in case the data transmission is not successful, the process keeps the current value of the BC unchanged to perform legacy EDCA (step 590) after which the process ends.

It is to be noted that the successful transmission may empty the corresponding AC queue, in which case the BC is no longer useful. In this context, embodiments provide that the modifying operation is applied further depending on whether there is still data in the traffic queue corresponding to the backoff counter or not.

In some embodiments, the queue storing the pre-empting data (to be sent) for an AC may be an alternative queue to the AC main queue. As an example, the alternate video (A_VI) and alternate voice (A_VO) transmit queues may share the same EDCAF as the VI and VO transmit queues, and be used to queue the pre-empting MSDU(s) for VI and VO, respectively. In that case also, the update of AC backoff is performed according to the filling status of the respective alternate queue: the modifying operation is applied depending on whether there is still data in the alternate traffic queue corresponding to the backoff counter or not (alternate queue empty as resulting from the pre-empted transmission). Indeed, there is no reason to degrade the pacing for the AC queue.

In the scenarios above, the backoff counter is reinitialized I reset at step 550, meaning a new initializing value is obtained. In variants, the modifying operation may merely consist in multiplying the current value of the backoff counter by a factor strictly higher than 1 , e.g., an integer factor. For example, the current value of the relevant BC is doubled.

In some embodiments, any transmission (initial one in the TXOP or pre-empting one in the pre-empted period 420) may provide more transmission time than needed by the targeted traffic class (e.g., primary AC). Multiple frame exchange comprising data from the primary AC (up to emptying the queue) and data from a secondary AC may be performed by the TXOP holder. This scenario looks like sharing an EDCA TXOP, wherein frames from a higher priority AC may be included when at least one frame from the primary AC has been transmitted and all frames from the primary AC have been transmitted (hence the primary AC is empty). To restore fairness due to this additional opportunity to send the secondary AC data, it may also be provided to perform step 550 for the secondary AC, in particular to reset the BC.

In this perspective, a dedicated communication method may be defined in a wireless network, that comprises, at a station (STA): using a first backoff counter to perform channel access and obtain a transmission opportunity (TXOP) to send first data associated with the first backoff counter; in case there is no more first data to be sent while the TXOP is on-going, sending second data associated with a second backoff counter, within the TXOP; and applying a modifying operation on a current value of the second backoff counter after having transmitted the second data and before starting using again the second backoff counter to perform channel access.

As above, the modification is preferably a reset (new initialization value randomly drawn from unchanged [0, CW] for the concerned AC) of the BC.

Step 550 of resetting the BC may be performed before or simultaneously to step 540 when it is triggered by the successful TXOP pre-emption, regardless of whether the transmission within the pre-empted period 420 is successful or not.

Once the STA has performed the data exchange 540 and the BC reset 550, it releases (step 560) the pre-empted TXOP, meaning the TXOP holder can take it back to use it. The release takes place at the end of the pre-empted period 420. If the length of the pre-empted period 420 is signalled in the PR frame 403, the TXOP holder only has to wait for the end of the pre-empted period 420. If it is not signalled, the TXOP holder may consider using a PIFS period after each transmission within the pre-empted period 420 (to try to take the TXOP back) whilst the preempting STA uses only SIFS periods as long as it wants to continue transmitting in the pre-empted period 420.

Once the medium has been released at step 560, the STA, in case it is not the AP, may transmit to its AP, an unsolicited Buffer Status Report (BSR) frame or the like to report a remaining queue size for the traffic queue corresponding to the data (traffic flow or AC or TID) transmitted during the pre-empted period. In other words, the STA updates the AP about the amount of data it has to transmit, in order for the AP to adjust the scheduled opportunities the AP can provide to the STA (e.g., in future trigger-based MU UL communications). The unsolicited BSR may be sent in a next EDCA transmission towards the AP.

In some embodiments where the TXOP holder is the AP, such signalling of the BSR may be provided to the AP within the pre-emption frame, i.e., in PR frame 403. In this case, the signalling is prior to the data transmission 540. However, as the BSR is usually used to report UL data (i.e., addressed to the AP), the AP will be aware of the success of the transmission 403 or not.

Figure 6a schematically illustrates a communication device 600, which may be any stations of radio network 100 of Figure 1 , configured to implement at least one embodiment of the present disclosure. The communication device 600 may preferably be a device such as a micro-computer, a workstation or a light portable device.

It should also be appreciated that the communication device is configured to operate in different modes (TXOP holder, TXOP share participant, source, intermediate, destination, first AP, other AP, stations associated with the first AP, stations associated with another AP, coordinator, coordinate, AP in an OBSS, STA in an OBSS, and so forth), depending on what role it performs in the current communication context. The communication device 600 comprises a communication bus 613 to which there are preferably connected: a central processing unit 601 , such as a processor, denoted CPU; a memory 603 for storing an executable code of methods or steps of the methods according to embodiments of the disclosure as well as the registers adapted to record variables and parameters necessary for implementing the methods; and at least one communication interface 602 connected to a wireless communication network, for example a communication network according to one of the IEEE 802.11 family of standards and/or Wireless-Fidelity (Wi-Fi) specifications, via transmitting and receiving antennas 604.

Preferably the communication bus 613 provides communication and interoperability between the various elements included in the communication device 600 or connected to it. The representation of the bus is not limiting and in particular the central processing unit 601 is operable to communicate instructions to any element of the communication device 600 directly or by means of another element of the communication device 600.

The executable code may be stored in a memory that may either be read only, a hard disk or on a removable digital medium such as for example a disk. According to an optional variant, the executable code of the programs can be received by means of the communication network, via the interface 602, in order to be stored in the memory of the communication device 600 before being executed.

In an embodiment, the device is a programmable apparatus which uses software to implement embodiments of the disclosure. However, alternatively, embodiments of the present disclosure may be implemented, totally or in partially, in hardware (for example, in the form of an Application Specific Integrated Circuit or ASIC).

Figure 6b is a block diagram schematically illustrating the architecture of the communication device 600, adapted to carry out, at least partially, some embodiments of the disclosure. As illustrated, device 600 comprises a physical (PHY) layer block 623, a MAC layer block 622, and an application layer block 621 .

The PHY layer block 623, here a plurality of 802.11 standardized PHY layer modules in case the device is a MLD (however a single PHY layer module may be contemplated when the device is single link), has the task of formatting, modulating on or demodulating from any 20MHz channel or composite channel or resource unit. The PHY layer thus sends or receives frames over the radio medium NETW, such as 802.11 frames. These frames may include frames to reserve and obtain a TXOP, data frames, Ack frames, PR frames as those shown in Figures 4 and 4a, and other conventional 802.11 frames.

The MAC layer block or controller 622 preferably comprises a MAC 802.11 layer 624 implementing conventional 802.1 1 MAC operations. It may comprise additional block 625 for carrying out, at least partially, embodiments of the disclosure. MAC layer block 622 may optionally be implemented in software, which software is loaded into RAM 603 and executed by CPU 601 . MAC 802.11 layer 624 may implement an Upper-MAC stack 624a along with one or more Lower- MAC modules 624b in case the device is an MLD. Of course, a single-link architecture is supported (whereas not illustrated here).

Preferably, additional block 625, referred to as “TXOP Pre-emption access” module, implements, in collaboration with MAC 802.11 layer 624, embodiments of the present disclosure to perform TXOP pre-emption operations to transmit, within an existing granted TXOP, pending data, such as data of latency sensitive traffic streams, as well as to perform fairness restoration, in particular by modifying when appropriate the current value of the backoff counter for which TXOP pre-emption has been gained and data have been sent in the pre-empted period 420. As an example, block 625 performs the operations of the methods illustrated in Figures 3 to 6.

On top of the Figure, application layer block 621 runs an application that generates and receives data packets, for example data packets such as a video stream. Application layer block 621 represents all the stack layers above MAC layer according to the ISO standardization.

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

Many further modifications and variations will suggest themselves to those versed in the art upon referring to the foregoing illustrative embodiments, which are given by way of example only and which are not intended to limit the scope of the disclosure, that being determined solely by the appended claims. In particular the different features 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

1. A communication method in a wireless network, comprising, at a station (STA): pre-empting a TXOP gained by a TXOP holder, for the STA to transmit data during a preempted period; and applying a modifying operation on a current value of a backoff counter after having preempted a TXOP, before starting using the backoff counter to perform channel access.

2. The method of Claim 1 , wherein the modifying operation includes drawing a new initializing value for the backoff counter.

3. The method of Claim 2, wherein in case the new initializing value is lower than the current value of the backoff counter, the current value of the backoff counter is kept unchanged to perform the next channel access.

4. The method of Claim 2, wherein the new initializing value is drawn from an unmodified contention window.

5. The method of Claim 1 , wherein the modifying operation includes multiplying the current value of the backoff counter by a factor strictly higher than 1 , e.g., an integer factor.

6. The method of Claim 1 , wherein applying the modifying operation on the current value of the backoff counter is in response to successfully pre-empting the TXOP.

7. The method of Claim 1 , wherein applying the modifying operation on the current value of the backoff counter is in response to successfully transmitting data within the pre-empted period.

8. The method of Claim 7, wherein the applying of the modifying operation is further dependent on whether there is still data in the traffic queue corresponding to the backoff counter or not.

9. The method of Claim 1 , wherein the STA includes a main traffic queue and a separate alternate traffic queue both associated with the backoff counter, the alternative traffic queue being used to store pre-empting data to be transmitted during pre-empted periods while the main traffic queue is used to store legacy data to be transmitted in TXOPs obtained through channel access, wherein the modifying operation is applied depending on whether there is still data in the alternate traffic queue corresponding to the backoff counter or not.

10. The method of Claim 1 , wherein the TXOP holder is a station different from the STA.

11 . The method of Claim 1 , wherein the STA is the TXOP holder that gained the TXOP for a given type of data and pre-empts its own TXOP to transmit another type of data.

12. The method of Claim 1 , further comprising, at the STA, transmitting to an access point (AP) to which it is associated, an unsolicited Buffer Status Report (BSR) frame to report a remaining queue size for the traffic queue corresponding to the data transmitted during the preempted period.

13. A communication method in a wireless network, comprising, at a station (STA): using a first backoff counter to perform channel access and obtain a transmission opportunity (TXOP) to send first data associated with the first backoff counter; in case there is no more first data to be sent while the TXOP is on-going, sending second data associated with a second backoff counter, within the TXOP; and applying a modifying operation on a current value of the second backoff counter after having transmitted the second data and before starting using the second backoff counter to perform channel access.

14. A wireless communication device comprising at least one microprocessor configured for carrying out the method of Claim 1 or 13.

15. A non-transitory computer-readable medium storing a program which, when executed by a microprocessor or computer system in a wireless device, causes the wireless device to perform the method of Claim 1 or 13.