The growing number of devices that require large amounts of data transmission presents a significant challenge to WiFi’s current capabilities. WiFi rises to this challenge with 802.11be, or WiFi 7. This new amendment proposes an evolution of unlicensed wireless connectivity as we know it. As the 802.11be standardization process solidifies, this article first provides an up-to-date summary of 802.11be’s basic features, demonstrating that multi-AP coordination is a must-have for critical and latency-sensitive applications. We then dive into the nitty-gritty of one of its most compelling implementations, coordinated beamforming, which our standards-matching simulations confirm reduces worst-case latency by nearly tenfold.
Back in 1943, psychologist Abraham Maslow published a study on the hierarchy of human needs, identifying four needs that must be met before a person’s talents and interests can be fully utilized. His theory can be vividly illustrated by a pyramid, from bottom to top: physiological needs, safety, belonging, and esteem. Today, one can "provocatively" add another layer to the bottom of Maslow's pyramid: WiFi. In addition to food, shelter, and clean water, wireless connectivity is also essential in our globalized society. While we would say that humans don't need the Internet more than air, the importance of WiFi is unquestionable. During the inconvenience of travel and isolation, many people use WiFi to stay in touch with loved ones, keep small businesses running through online orders, or stay healthy through online yoga classes. After all, this article would be almost impossible to write without WiFi, and you are likely using WiFi as you read it.
Billions of people use WiFi every day, carrying the majority of the world’s data traffic in an ever-expanding variety of applications. By 2023, there will be nearly 628 million public WiFi hotspots, with one in ten equipped with WiFi 6 based on the IEEE 802.11ax specification. As WiFi grows in popularity and functionality, so will the demand for wireless services. In addition to 8K displays and VR, more homes will use smart appliances, which will become a dense environment with many devices connected simultaneously. Businesses will significantly increase the amount of data collected on their premises, improving manufacturing processes and increasing productivity. Importantly, this cross-plant communication can provide very low latency to enable machinery synchronization and real-time control. Real-time video will account for a large portion of global IP traffic, and after the pandemic, high-quality video conferencing will be widely used in fields such as work, education, and healthcare.
Our need for high speed and high reliability has driven the development of the next generation of WiFi 7 based on IEEE 802.11be Extremely High Throughput (EHT). Since 802.11be was introduced to our research community, regulatory, certification, and standardization bodies have done a lot of work. The Federal Communications Commission (FCC) has made new spectrum in the 6GHz frequency available for unlicensed use. The WiFi Alliance, a global network of companies that drives the adoption and development of WiFi through certification, is expected to soon provide global interoperability certification for WiFi 6 devices to operate in such a new frequency band. In the meantime, key experts are conducting a virtual IEEE meeting to determine the building blocks of the 802.11be standard.
In this article, we start with an update on the current state of WiFi and predict how it will evolve in the future. We then take a detailed look at the latest specific decisions on the technical features that will be adopted in the 802.11be amendment, as well as the new expected development timeline. We also discuss one of the most attractive factors for improving network efficiency, lower latency, and improved reliability to complement the increase in peak throughput: multi-access point (AP) coordinated beamforming (CBF). In particular, we shed light on the details of its potential implementation and share standards-compliant simulation results that quantify the latency gains it achieves in a realistic digital enterprise setting.
Coping with more stringent requirements in high-density scenarios is one of the most challenging goals that WiFi must achieve. The most advanced WiFi 6 based on IEEE 802.11ax addresses congestion by improving network efficiency and battery consumption through features such as orthogonal frequency division multiple access (OFDMA) and uplink and downlink multi-user MIMO. With 802.11ax not yet finalized until the end of 2020, WiFi stakeholders have begun to focus on two further improvements to WiFi 6. The first is WiFi 6E, which is currently being opened up by governments around the world for unlicensed use. The second will be a new 802.11be amendment that may be certified as WiFi 7.
A. WiFi 6E: A new WiFi track
For more than 20 years, WiFi has operated in two frequency bands, 2.4 and 5 GHz. In April 2020, the FCC cleared the way for a third band: 5.925–7.125 GHz. This added spectrum, called the 6 GHz band, has nearly four times the available bandwidth. In addition to more available channels, a key difference of the newly opened frequencies is their shorter propagation range, which may be particularly suitable for providing basic service set (BSS) isolation in dense and challenging environments such as transportation hubs, stadiums, and shopping districts. Under rules established to protect existing services, the new 6 GHz band will be accessed by licensed devices. Among them, outdoor use will have a mandatory competition-based agreement to limit total transmission power and power spectral density to avoid inefficient use of narrow channels.
While the FCC's decision may put the United States ahead in the 6GHz market, other regions including Europe and Asia Pacific are also exploring unlicensed access to the band. In the meantime, WiFi 6 is ready to take advantage of the 6 GHz spectrum available worldwide, and devices equipped with the chips and radios needed to operate in the new band will receive a "6E" designation, where E stands for "extended." The WiFi Alliance plans to launch WiFi 6E certification in early 2021, with more than 300 million compliant devices expected to be available in the same year. It's worth noting that since only 6E devices will initially be able to work in the 6GHz band, they will at least be available in their original, low-interference settings.
B. WiFi 7: (Not Just) Very High Throughput
In fact, 802.11be’s extremely high throughput will far exceed the high peak data rates. It is very certain that WiFi 7 is expected to support at least 30 Gbps per AP, about four times that of WiFi 6, while ensuring backward compatibility and coexistence with legacy devices in the 2.4, 5, and 6 GHz unlicensed bands. However, the 802.11be Task Group (TG) also recognizes the need and is working on lower latency and higher reliability to enable time-sensitive networking (TSN) use cases. The former is seen as an enabler for real-time applications (including augmented and virtual reality, gaming, and cloud computing), requiring latency to be reduced to less than 5 milliseconds. The latter is critical for next-generation factories and enterprises, where WiFi may need to guarantee higher reliability to replace certain wired communications.
To accelerate the development and commercialization of WiFi 7 (whose timeline is shown in Figure 1), the 802.11be TG deviates from the traditional single-phase development cycle and identifies two phases. The first phase focuses on a set of high-priority features based on their gain/complexity ratio, standardization and implementation time, and related and market needs. This is described in detail in the next section.
Figure 1: Illustration of the current WiFi 7 standardization, certification, and commercialization timeline.
Introduction to the future of WIFI 7
At the time of writing, the 802.11be TG is actively defining the basic functional operations that will be included in the standard. This information is collected in the Specification Framework Document (SFD), from which the draft standard will be derived. We mainly focus on its subsequent updates.
Figure 2: Illustration of the PSR framework.
A. First Release Features
As shown in Figure 1, the first release (R1) features are expected to reach mature specifications in draft 1.0 by the time they are due in May 2021, with the possibility of further extensions and improvements to them by the release of draft 2 in March 2022. These include:
(1) Multi-link operation:
802.11be targets efficient operation in all available frequency bands (i.e., 2.4, 5, and 6 GHz) for load balancing, multi-band aggregation, and simultaneous downlink/uplink transmissions. In 802.11be, a multi-link device (MLD) is defined as a device with multiple subordinate APs or STAs and a single MAC service access point (SAP) to the above-mentioned logical link control (LLC) layer. A MAC address that uniquely identifies the MLD management entity is also introduced. The relevant functions of multi-link control and operation are summarized as follows:
-
Multi-link discovery and setup: MLD has the ability to dynamically update its frame exchange on each pair of links simultaneously. In addition, each individual AP/STA can also provide information about the operating parameters of other affiliated AP/STAs within the same MLD.
-
Traffic Link Mapping: In a multi-link setup, all traffic identifiers (TIDs) used to classify frames according to their Quality of Service (QoS) are mapped to all setup links. Updates to this mapping may be made subsequently by any associated MLD. In addition, the receiving MLD will utilize a single reordering buffer for QoS data frames of the same TID transmitted over multiple links.
-
Channel access and power saving: Each AP/STA of MLD performs independent channel access through its link and maintains its own power state. To facilitate efficient STA power management, the AP can also use the enabled link to carry indications of buffered data for transmission on other links.
(2) Low-complexity AP coordination:
802.11be will support multi-AP coordination, with the AP implementing its functions in beacon frames/management frames. Coordinated spatial reuse (CSR) is a low-complexity implementation that can be included in R1. In CSR, a shared AP that has obtained a transmission opportunity (TXOP) can trigger one or more other shared APs to perform synchronized transmissions with appropriate power control and link adaptation. Compared to the spatial reuse scheme available in 802.11ax, this coordination will create more spatial reuse opportunities and reduce the number of conflicts.
(3) Direct enhancements to 802.11ax:
The 802.11be TG will also specify some upgrades to the current 802.11ax standard. These include:
-
Support for 320 MHz transmissions, doubling the 160 MHz transmissions of 802.11ax.
-
Using higher modulation orders, 4096 QAM can be supported, while only 1024-QAM is supported in 802.11ax, and there is a strict -38 dB requirement for the error vector magnitude (EVM) of the transmitter.
-
Each STA is allocated multiple resource units, namely OFMDA. Its flexibility can improve spectrum utilization.
B. Second Edition Features
Although the R2 features will be formalized in Draft 3.0 and Draft 4.0 in November 2022 and November 2023 respectively, the 802.11be TG has already started its work and made significant progress in SFD. The main features are as follows:
1.MIMO enhancement:
802.11be doubles the maximum number of supported single-user MIMO (SU-MIMO) and multi-user MIMO (MU-MIMO) spatial streams to 16, thereby increasing capacity. In the case of MUMIMO, the 802.11be TG agreed to limit the maximum number of spatial multiplexing STAs and spatial streams per STA to 8 and 4, respectively. The above restrictions help control MIMO precoder complexity and channel state information (CSI) overhead. Research on implicit CSI detection is currently underway, which can be used as an optional mode to further suppress this overhead.
2. Hybrid Automatic Repeat Request (HARQ):
R2 may introduce HARQ. Instead of discarding erroneous information, the device tries to soft-combine it with the retransmission unit to increase the probability of correct decoding. Although SFD does not include any HARQ-related processes at the time of writing, the 802.11be TG has evaluated different HARQ units MAC Protocol Data Unit (MPDU) or PHY codewords and evaluated the performance/complexity trade-offs.
3. Low latency operation:
Given the commercial appeal of TSN, SFD will also collect protocol enhancements specifically designed to reduce worst-case latency and significantly improve reliability. It is conceivable that such a solution may rely on multi-link operation, providing different QoS for each link, or rely on AP coordination to achieve more aggressive spectrum reuse and fewer harmful conflicts.
4. Advanced AP coordination
: In order to fully tap the potential of multi-AP coordination, 802.11be TG agreed to support the following three solutions:
-
Coordinated OFDMA: In 802.11be, an AP that obtains a TXOP will be able to share its frequency resources in multiples of a 20 MHz channel with a set of neighboring APs. For efficiency, a shared AP can request neighboring APs to report their resource requirements.
-
Single-user and multi-user joint transmission: Sending data to its connected STAs requires the AP to bind its phase synchronization error and timing offset. When considering the reasonable range of these deviations, it is found that joint transmission can bring gains under the premise of sufficient backhaul. Since the cooperative AP needs CSI from related and non-related STAs, 802.11be will define a joint multi-AP detection scheme. In this way, the AP will send its detection frame at the same time, and the addressed STA will transmit the CSI feedback of all APs.
-
Coordinated Beamforming: This technique exploits the ability of modern multi-antenna APs to spatially multiplex their STAs while jointly placing radiation nulls to or from neighboring non-associated STAs. While the CSI required to control the radiation nulls can be obtained through the joint multi-AP sounding scheme described above, CBF can also take advantage of the simpler sequential sounding procedure that will be part of 802.11be. In addition, CBF does not require joint data processing because each STA sends or receives data to or from a single AP, thus significantly reducing the number of joint transmissions required for the backhaul. This is because CBF can provide significant throughput and latency enhancements while maintaining complexity, which we will explore further in the next section.
Enhanced spatial multiplexing through multi-AP coordinated beamforming
There is a consensus that building reliability and low latency features on top of 802.11ax will facilitate backward compatibility, product certification, and market adoption. To this end, parameterized spatial reuse (PSR) in 802.11ax is an attractive module because it allows dynamic collaboration between devices in different BSSs. Next, we will introduce the PSR framework, discuss its advantages and disadvantages, and explain how to extend it through multi-AP coordination to suppress latency and improve reliability in 802.11be.
A. Parameterized Spatial Multiplexing in 802.11ax
In PSR, an AP that needs to perform uplink reception can offer a TXOP to an overlapping BSS (OBSS) via a trigger frame. In its basic form, a trigger frame can be considered a scheduling grant that provides information and timing for subsequent uplink transmissions. When PSR is enabled, the AP can use the trigger frame to invite OBSS devices to reuse the spectrum at the same time as its uplink reception, provided they meet certain interference conditions.
To provide a more detailed description of the PSR framework, let us look at Figure 2(a) with an example of two BSSs, where:
-
BSS1 consists of AP1, STA11, and STA12; BSS2 includes AP2, STA21, STA22, and STA23.
-
Figure 2(b) shows how AP1 initiates PSR processing by sending a trigger frame after gaining channel access. This trigger frame has a dual function:
-
Transmitting synchronization and scheduling information required for uplink transmission of its associated STA11 and STA12; and
-
The spatial reuse opportunity is advertised to the OBSS device, which spans the subsequent uplink data received by AP1.
To ensure that transmissions utilizing spatial multiplexing opportunities do not affect AP1,’s uplink data reception, the trigger frame contains a PSR field. This field contains the following information: i) the maximum interference level that AP1, can receive without affecting its uplink reception; ii) AP1,’s transmit power to facilitate interference calculations. Upon receiving the trigger frame, OBSS devices measure their received power level and, based on the information provided in the PSR field, determine whether they can access the medium and at what transmit power.
In the example of Figure 2(b), STA21, STA22, and STA23 all have uplink data to send. However, only STA21 and STA22 are able to independently determine that they can compete for the medium. Unfortunately, STA23 is unable to compete for channel access because it is close to AP1 and cannot meet the interference conditions set by the latter. The end result is that STA21 first accesses the channel to send its short packet, ensuring that the corresponding acknowledgment (ACK) frame is received within the duration of the uplink transmission triggered by AP1. As long as this duration allows, STA22 will also have the opportunity to re-compete for the channel and transmit.
1. Advantages of PSR:
Overall, thanks to the PRS framework, APs and STAs can gain channel access, which improves spatial reuse and, in turn:
-
Increases network throughput because it allows more concurrent transfers;
-
Increased STA file throughput because STA spends less time in contention; importantly,
-
Reduced latency, because STAs with time-sensitive short file traffic may not need to wait until the broadband STA terminates its long transmission. This is the case for STA21 and STA22 in Figure 2.
2. PSR Challenges:
Although the PSR framework allows for greater spatial reuse, two challenges were discovered during the 802.11be research:
-
Devices that take advantage of spatial reuse opportunities must reduce their transmit power to limit the interference generated. In some cases, for STA21, STA22 in Figure 2, this translates into reduced throughput. In other cases, for STA23, the device cannot even access the spatial reuse opportunity because its maximum allowed transmit power is insufficient to reach its receiver.
-
Devices that exploit spatial multiplexing opportunities are unaware of and cannot control the interference perceived by their respective receivers. In Figure 2, this means that if STA21, STA22 is close to AP2, the uplink transmission from STA21, STA22 to AP2 may fail because AP2 will receive a non-negligible amount of interference from STA21, STA22.
The above two shortcomings hinder the effectiveness of existing PSR frameworks in various settings, including high-density scenarios or scenarios where devices handle latency-sensitive data traffic and cannot afford transmission failures or excessive channel access latency.
Figure 3: Illustration of the coordinated beamforming protocol.
B. Coordinated Beamforming in 802.11be
802.11be aims to take the existing spatial reuse capabilities to a whole new level through CBF, i.e., by letting cooperative APs suppress incoming OBS interference in the spatial domain. Recent experimental studies have shown that a four-antenna AP serving one STA is able to suppress up to 10 db of interference to neighboring links compared to a single-antenna system. Based on these results, we now detail an illustrative protocol that implements CBF by smoothly building on the PSR framework.
Let us look at the uplink transmission scheme of Figure 3(a). The setup is similar to Figure 2(a), but AP1 and AP2 are now equipped with eight antennas. The proposed CBF protocol has three phases, of which the first two phases are common to the CBF and joint transmission implementations currently discussed in 802.11be, shown in Figure 3(b) and described as follows.
1. Multi-AP coordination:
In this phase, two or more collaborative APs exchange control frames for two purposes:
-
Establishment and maintenance of coordination sets: In order for CBF to be effective, APs need to communicate with OBSS STAs, such as obtaining the necessary CSI to place radiation nulls at specific spatial locations. To this end, an inter-BSS coordination set is defined between the cooperating APs, which must contain the IDs of all APs and STAs participating in the CBF transmission. These IDs can be kept in memory by all relevant devices without discarding the relevant frames generated by the OBSS devices included in their coordination set as in the traditional way. Once defined, the inter-BSS coordination set can be updated in a semi-static manner (i.e., after tens or hundreds of TXOPs).
-
Dynamic coordination of subsequent spatial reuse opportunities: Once AP1 obtains a TXOP, it needs to announce the incoming uplink triggered transmission and, together with the devices in its coordination set, determine which STAs will participate in the subsequent CSI acquisition and data communication phases. In the example of Figure 3(b), AP2 replies to the dynamic coordination frame sent by AP1, indicating which of its STAs will benefit most from being granted a secure spatial reuse opportunity, e.g., STA21 and STA22.
2. CSI Acquisition:
At this stage, due to previous coordination, both AP1 and AP2 acquire CSI only from relevant BSS internal and OBS devices. This CSI is necessary in order to design filters for spatial multiplexing and bidirectional interference suppression in the subsequent communication stage. Importantly, as OBSS devices are addressed to acquire CSI, they realize that the OBSS AP will soon provide them with spatial multiplexing opportunities with more favorable channel access conditions. Since no new specific signaling is required to trigger data communication, the 802.11ax trigger frame can be used for this purpose. This brings obvious benefits to traditional STAs, which can continue to apply the traditional PSR framework of 802.11ax in a seamless manner.
3. Data communication:
The implementation of the first two phases addresses the two fundamental challenges of the 802.11ax PSR framework highlighted in the previous section, making spatially multiplexed transmissions from STA21, STA22, and STA23 more likely to succeed under adverse conditions. This is because:
-
STA21, STA22 and even STA23 are more likely to find spatial reuse opportunities and use their maximum transmission power. This is thanks to the spatial interference mitigation performed by AP1, which helps to publish relaxed messages about the channel access conditions of the associated OBSS devices.
-
AP2 is now able to suppress incoming interference generated by STA11 and STA12 while receiving uplink transmissions from STA21, STA22, and STA2.
802.11BE Coordinated Beamforming Performance
We now quantify the latency enhancement provided by the CBF scheme described in the previous section. With this goal in mind, we consider the deployment of 2 roof-mounted APs, each equipped with 8 antennas and 24 STAs, which are evenly distributed in a 35m*20m*3m indoor space. Of these 24 STAs, 16 STAs generate uplink broadband traffic and the remaining 8 STAs generate uplink latency-sensitive AR traffic. Since our main goal is to guarantee the on-time delivery of AR services, the AP that grants spatial reuse opportunities will suppress interference from neighboring AR STAs, which generate the strongest interference, usually corresponding to the STA located at the closest location. The results in this section are the results of system-level simulations in a complex standard, and its basic settings are detailed in Table 1. Interested readers can find the full set of simulation parameters there.
Figure 4 shows the median values experienced by an AR STA under three different settings: 5%, 1%, and 0.01% worst-case MAC layer latency:
Table 1: System-level simulation parameters
Figure 4: Median and worst-case latency (ms) experienced by an AR STA. Three systems were evaluated: 1) IEEE 802.11ax without spatial reuse, 2) IEEE 802.11ax with PSR, and 3) IEEE 802.11be with CBF.
-
IEEE 802.11ax device setup without spatial multiplexing: The results in Figure 4 show that IEEE 802.11 based systems may be able to provide low latency but struggle to maintain consistent performance in the worst case. In fact, we can observe that latency remains below 3ms for about 50% of the considered scenarios, but increases significantly to over 200ms in the 0.01% worst case scenario. This is mainly due to the combined effects of the random channel access mechanism and collisions leading to retransmissions.
-
Setup with IEEE 802.11ax devices supporting PSR: Figure 4 illustrates that the implementation of PSR does not help to significantly reduce the worst-case latency. This is because, similar to STA23 in Figure 2, the AR STAs are not far enough from their neighboring APs in the considered dense scenario. This prevents these latency-sensitive STAs from discovering spatial reuse opportunities because the channel access conditions they need to comply with to prevent harmful interference are too stringent, as detailed in Section 4.2.
-
Equipment setup implementing the IEEE 802.11be CBF scheme described in the previous section: The results in Figure 4 illustrate how the proposed scheme significantly reduces the worst-case latency compared to other IEEE 802.11ax systems. In fact, we can observe that the system with multi-AP coordination capability reduces the worst-case latency of 0.01% by a factor of 9 relative to the system with PSR capability. This significant performance enhancement is a direct result of i) the large number of spatial multiplexing opportunities discovered by the AR STAs due to their relaxed channel access conditions, and ii) the OBSS interference mitigation provided in the spatial domain, which maximizes the chances of performing a successful data transmission.
It should be noted that the throughput of wideband STA remains roughly the same for the three systems under evaluation.
The next generation of WiFi will open access to gigabit, high-reliability, and low-latency communications, reshaping manufacturing and social connectivity through digital enhancement. In this article, we detail the steps taken by IEEE 802.11be to achieve WiFi 7, the latest protocols for its technical features, and the latest timeline. We illustrate the importance of spatial reuse through multi-AP coordinated beamforming, shared implementation details, and standard-compliant simulations. In the future, further research is needed to incorporate these technologies into time-sensitive network protocols to make wireless the new wired network for our homes and industries.
Note: The original article is "IEEE 802.11be: Wi-Fi 7 Strikes Back", the original author is: Adrian Garcia-Rodriguez, David Lopez-P ´ ´erez, Lorenzo Galati-Giordano, and Giovanni Geraci, special thanks!
Reply "WiFi 7 original text"
in the official account
to get the original text download link~
*Disclaimer: This article is originally written by the author. The content of the article is the author's personal opinion. Semiconductor Industry Observer reprints it only to convey a different point of view. It does not mean that Semiconductor Industry Observer agrees or supports this point of view. If you have any objections, please contact Semiconductor Industry Observer.
Today is the 2930th content shared by "Semiconductor Industry Observer" for you, welcome to follow.
Semiconductor Industry Observation
"
The first vertical media in semiconductor industry
"
Real-time professional original depth
Scan the QR code
, reply to the keywords below, and read more
Wafers|ICs|Equipment
|Automotive Chips|Storage|TSMC|AI|Packaging
Reply
Submit your article
and read "How to become a member of "Semiconductor Industry Observer""
Reply
Search
and you can easily find other articles that interest you!