What awaits us in Wi-Fi 7 IEEE 802.11be?
In September 2020, we will be celebrating the 30th anniversary of the IEEE 802.11 project, which has had a significant impact on our lives. Currently, Wi-Fi technology, defined by the IEEE 802.11 family of standards, is the most popular wireless technology used to connect to the Internet, with Wi-Fi carrying over half of user traffic. While cellular technology rebrands every decade, such as changing the name from 4G to 5G, for Wi-Fi users, the increase in data speeds, as well as the introduction of new services and new features, are almost invisible. Few customers care about the "n", "ac", or "ax" that follows "802.11" on equipment boxes. But that doesn't mean Wi-Fi isn't evolving.
One piece of evidence for the evolution of Wi-Fi is the dramatic increase in nominal data rates, from 2 Mbps in the 1997 version to almost 10 Gbps in the latest 802.11ax standard, also known as Wi-Fi 6. Modern Wi-Fi reaches such performance gains due to faster signal-code designs, wider channels, and the use of MIMO technologies.
In addition to the mainstream of high-speed wireless LANs, the evolution of Wi-Fi includes several niche projects. For example, Wi-Fi HaLow (802.11ah) was an attempt to bring Wi-Fi into the wireless Internet of Things market. Millimeter-wave (802.11ad/ay) Wi-Fi supports nominal data rates up to 275Gbps, though over very short distances.
New applications and services related to high-definition video streaming, virtual and augmented reality, gaming, remote office and cloud computing, as well as the need to support a large number of users with high traffic on wireless networks, require high performance.
Wi-Fi 7 Goals
In May 2019, Subgroup BE (TGbe) of the 802.11 Working Group of the Local and Metropolitan Networking Standards Committee began work on a new addition to the Wi-Fi standard that will increase the nominal throughput to more than 40 Gbps in a single "typical" frequency channel. for Wi-Fi band <= 7 GHz. Although many documents mention "maximum throughput of at least 30 Gb / s", the new physical layer protocol will provide nominal speeds in excess of 40 Gb / s.
Another important direction in the development of Wi-Fi 7 is the support for real-time applications (games, virtual and augmented reality, robot control). It is noteworthy that although Wi-Fi serves audio and video traffic in a special way, for a long time it was believed that providing guaranteed low delays (units of milliseconds), also known as Time-Sensitive Networking, at the standard level, is fundamentally impossible in Wi-Fi networks. In November 2017, our team from IPTP RAS and NRU HSE (do not take it for PR) made a corresponding proposal in the IEEE 802.11 group. The proposal generated a lot of interest and an ad hoc subgroup was launched in July 2018 to look into the matter further. Since real-time application support requires both high nominal data rates and increased link-layer functionality, the 802.11 working group decided to develop methods to support real-time applications within Wi-Fi 7.
An important issue related to Wi-Fi 7 is its coexistence with cellular network technologies (4G/5G) being developed by 3GPP and operating in the same unlicensed frequency bands. We are talking about LTE-LAA/NR-U. To study the problems associated with the coexistence of Wi-Fi and cellular networks, IEEE 802.11 launched the Coexisting Standing Committee (Coex SC). Despite numerous meetings and even a joint workshop of 3GPP and IEEE 802.11 participants in July 2019 in Vienna, technical solutions have not yet been approved. A possible explanation for this fruitless activity is that both IEEE 802 and 3GPP are unwilling to change their own technologies to bring them into line with the other. Thus, at the moment it is not clear whether the discussions within the Coex SC will affect the Wi-Fi 7 standard.
While the Wi-Fi 7 development process is in its very early stages, there have been about 500 new feature proposals to date for the future of Wi-Fi 7, also known as IEEE 802.11be. Most of the ideas are only being discussed in the be subgroup and have not yet been decided on. Other ideas have recently been approved. Below it will be clearly indicated which proposals are approved and which are only under discussion.
It was originally planned that the development of the main new mechanisms would be completed by March 2021. The final version of the standard is expected by early 2024. In January 2020, concerns were expressed in the 11be subgroup about whether development would stay on schedule at the current pace of work. To speed up the standard development process, the subgroup agreed to select a small set of high priority features that could be released by 2021 (Release 1) and leave the rest to Release 2. The high priority features should provide major performance gains and include support for 320 MHz, 4K- QAM, obvious OFDMA improvements from Wi-Fi 6, MU-MIMO with 16 streams.
Due to the coronavirus, the group is currently not meeting face-to-face, but regularly holds teleconferences. Thus, development slowed down somewhat, but did not stop.
Consider the main innovations of Wi-Fi 7:
1. The new physical layer protocol is an evolution of the Wi-Fi 6 protocol with a 2x increase in bandwidth to 320 MHz, a 2x increase in the number of MU-MIMO spatial streams, which increases the nominal throughput by 2x2 = 4 times. Wi-Fi 7 is also starting to use 4K-QAM modulation, which adds another 20% to the nominal throughput. Thus, Wi-Fi 7 will provide a nominal data transfer rate of 2x2x1.2 = 4.8 times higher than Wi-Fi 6: Wi-Fi 7 maximum nominal throughput is 9.6 Gbps x 4.8 = 46 Gbps. In addition, there will be a revolutionary change in the physical layer protocol to ensure compatibility with future versions of Wi-Fi, but it will remain invisible to users.
2. Changing the channel access method to support real-time applications will be carried out taking into account the experience of IEEE 802 TSN for wired networks. Ongoing discussions in the standardization committee are related to the random backoff procedure for channel access, traffic service categories and, accordingly, separate queues for real-time traffic, as well as packet service policies.
3. Introduced in Wi-Fi 6 (802.11ax), OFDMA - a time and frequency division channel access method (similar to that used in 4G and 5G networks) - provides new opportunities for optimal resource allocation. However, in 11ax OFDMA is not flexible enough. First, it allows the access point to allocate only one resource block of a predetermined size to the client device. Secondly, it does not support direct transmission between client stations. Both disadvantages reduce the spectral efficiency. In addition, the lack of flexibility inherited from Wi-Fi 6 OFDMA degrades performance in dense networks and increases latency, which is critical for real-time applications. 11be will solve these OFDMA problems.
4. One of the claimed revolutionary changes in Wi-Fi 7 is the built-in support for the simultaneous use of multiple parallel connections on different frequencies, which is very useful for both huge data transfer rates and extremely low latency. Although modern chipsets can already use multiple connections at the same time, for example, in the 2.4 and 5 GHz bands, these connections are independent, which limits the effectiveness of such an operation. In 11be, a level of synchronization between channels will be found that allows efficient use of channel resources and will entail significant changes in the rules of the channel access protocol.
5. The use of very wide channels and a large number of spatial streams leads to the problem of high overhead associated with the channel state estimation procedure required for MIMO and OFDMA. These overheads negate any benefit from higher nominal data rates. It is expected that the channel state estimation procedure will be revised.
6. In the context of Wi-Fi 7, the standards committee is discussing the use of some "advanced" data transfer methods. In theory, these techniques improve spectral efficiency in the case of retransmission attempts, as well as simultaneous transmissions in the same or opposite directions. We are talking about the hybrid automatic repetition request (HARQ), currently used in cellular networks, the full-duplex mode and non-orthogonal multiple access (NOMA). These methods have been well studied in the literature in theory, but it is not yet clear whether the performance gain they provide will outweigh the effort put into their implementation.
● The use of HARQ is complicated by the following problem. In Wi-Fi, packets are concatenated to reduce overhead. In current versions of Wi-Fi, the delivery of each packet inside the glued is confirmed and, if no acknowledgment is received, the transmission of the packet is repeated by the methods of the channel access protocol. HARQ transfers retries from the data link to the physical layer, where there are no more packets, but there are code words, and the code word boundaries do not coincide with the packet boundary. This desynchronization complicates the implementation of HARQ in Wi-Fi.
● As far as Full-Duplex is concerned, neither cellular networks nor Wi-Fi networks can currently transmit data to and from the access point (base station) on the same frequency channel at the same time. From a technical point of view, this is due to the large difference in the power of the transmitted and received signal. Although there are prototypes that combine digital and analog subtraction of the transmitted signal from the received signal, capable of receiving a Wi-Fi signal during their transmission, the gain that they can give in practice may be negligible due to the fact that at any given time the downstream is not is equal to the ascending one (on average, the descending one is much larger “in the hospital”). At the same time, such two-way transmission will significantly complicate the protocol.
● If multi-stream transmission using MIMO requires multiple sender and receiver antennas, in case of non-orthogonal access, the access point can simultaneously transmit data to two receivers from the same antenna. Various non-orthogonal access options are included in the latest 5G specifications. The prototype of NOMA Wi-Fi was first created in 2018 at the IPTP RAS (again, do not take it as a PR). It showed a performance gain of 30-40%. The advantage of the developed technology is its backward compatibility: one of the two recipients may be an outdated device that does not support Wi-Fi 7. In general, the problem of backward compatibility is very important, since devices of different generations can work simultaneously in a Wi-Fi network. Currently, several teams around the world are analyzing the effectiveness of the combined use of NOMA and MU-MIMO, the results of which will determine the future fate of the approach. We also continue to work on the prototype: its next version will be presented at the IEEE INFOCOM conference in July 2020.
7. Finally, another important innovation, but with an unclear fate, is the coordinated work of access points. Although many vendors have their own centralized controllers for enterprise Wi-Fi networks, the capabilities of such controllers have generally been limited to configuring long-term parameters and channel selection. The Standards Committee is discussing closer cooperation between neighboring access points, which includes coordinated transmission scheduling, beamforming, and even distributed MIMO systems. Some of the considered approaches use sequential interference suppression (roughly the same as in NOMA). Although approaches for 11be coordination have not yet been developed, there is no doubt that the standard will allow access points from different manufacturers to coordinate transmission schedules with each other in order to reduce mutual interference. As for other, more complex approaches (for example, distributed MU-MIMO), it will be more difficult to implement them into the standard, although some members of the group are determined to do this in Release 2. Regardless of the outcome, the fate of access point coordination methods is vague. Even being included in the standard, they may not reach the market. This has happened before when trying to clean up Wi-Fi transmissions with solutions like HCCA (11e) and HCCA TXOP Negotiation (11be).
In summary, it seems that most of the proposals associated with the first five groups will become part of Wi-Fi 7, while the proposals associated with the last two groups require significant additional research to prove their effectiveness.