What is new in 802.11ax
802.11ax promises a lateral progression of the WiFi standard. Until now, all new WiFi specifications were focused on improving the peak data rates. 802.11ax takes a side step and promises higher efficiency rather than higher peak rates. It promises to improve the total throughput in a cell and deliver more data to the clients rather than try to improve the speed at which the data is transmitted. 802.11ax features will be available in 2.4 GHz band as well (as opposed to 802.11ac which was available only in 5GHz).
To achieve this goal it introduces a few new features.
- Orthogonal frequency-division multiple access (OFDMA)
The AP can break up the available spectrum into smaller chunks (called RUs or ‘resource units’) so that it can transmit data to multiple clients simultaneously. Each client simply needs to filter the signals from the RU assigned to it. These RUs are also used for uplink transmissions. See additional details here.
- Multiuser uplink
In addition to the MU-MIMO on downlink (AP to clients) introduced with 802.11ac Wave 2; 802.11ax allows for the simultaneous transmissions from multiple clients to the AP. This requires the AP to schedule and synchronize these transmissions from the clients.
- Spatial reuse
For more efficient use of the spectrum, 802.11ax allows a device (AP or client) to transmit a frame even while another frame is in flight on the same channel as long as the other frame is not for its own BSS. The 802.11ax device does this by using a variable clear channel assessment threshold thereby allowing itself to transmit even when another BSS’s (low RSSI) frame is in the air. The device simultaneously reduces its transmit power so as to not interfere with the other BSS frame. Additional details here.
Apart from these, there are numerous other improvements to the existing features of 802.11 standard.
- Uses ¼ the subcarrier spacing – reduced from 312.5 kHz to 78.125 kHz – increases the number of subcarriers 4 times.
- Allows for 1024 QAM (as compared to up to 256 in 802.11ac) – allowing for higher modulation rates.
- Larger guard intervals – allows for larger distance by improving the management of fading and inter symbol interference
- Increases number of simultaneous downlink MU-MIMO group from 4 to 8 – allows for more simultaneous downlink transmissions
- Introduces target wake time – allows low data clients to sleep longer
This is great; should I replace my APs with 802.11ax APs?
While the above list of improvements is quite impressive, there are a few practical considerations that must be thought through before a rip-and-replace trigger.
Note that many of the above features (OFDMA, Multiuser Uplink, smaller subcarrier spacing, 1024 QAM, longer OFDM symbols, etc.) require both the AP and the client to be 802.11ax compliant.
If you have a bunch of pre 802.11ax clients, you will not be able to get any meaningful improvement from just replacing the APs. While it is relatively easier to control the client mix in a corporate environment; legacy devices will continue to force the 802.11ax deployments to legacy rates in places like hotspots, retail and home.
MU-MIMO, while already introduced into 802.11ac Wave 2, has not been seen to deliver on its promise because of the complex math and sounding overhead required to make it work. See a very insightful video by Devin Akin here.
It still remains to be tested how much benefit newer features of 802.11ax will provide in real life scenarios. My friend and colleague Hemant Chaskar explains these here.
OFDMA and Multiuser uplink will require intelligent grouping of appropriate of devices (taking into consideration things like client’s radio environment, traffic type/priority, QoS requirements, client state, etc.) to be able to derive the benefits of these features. For example, devices grouped in a multiuser uplink group should have enough queued data to be able to fill the PPDU, the AP needs to instruct the clients (and the clients need to honor) to adjust their power to equalize the signal strength from different clients at the AP.
Another example: spatial reuse requires reduction of transmit power in conjunction with increasing the CCA thresholds. This means the client or AP has to reduce the MCS of the transmitted packet and could result in a longer occupancy of the channel, effectively causing other devices to back off.
1024 QAM and narrower subcarriers will require higher SNRs and devices will require radios with tighter quality and higher sensitivity.
So, what should we expect?
It is a given that 802.11ax APs will start to replace 802.11ac APs on the shelves starting fall or late 2018. 802.11ax client radios will also start appearing on high-end devices in the same timeframe. However, for the near term, i.e. well into 2019, legacy client radios will continue to dominate the scene; both because of still operational legacy devices and lower-end devices that keep shipping with pre-802.11ax clients.
On the AP side, it is unlikely that all the features of 802.11ax will be supported from the beginning. Even chipset vendors are likely to add features in stages (like what happened for both 802.11ac Wave 1 and Wave 2). In this case AP and client capability mismatch will also force legacy operation. Some of these features could be added via software upgrades but some will require new hardware.
Early 802.11ax APs are also unlikely to have figured out optional algorithms for features like spatial reuse thresholds and multiuser uplink client grouping. Most of these algorithms could be software upgraded.
Also, we need to note that the 802.11ax is not yet ratified and the final approval is expected only towards the late 2019. It is, however, unlikely to have major changes from the current draft.
The best approach to 802.11ax would be as follows:
- If you have pre-802.11ac APs that need immediate refresh and you will also be refreshing your current WiFi devices with higher end devices in late 2018 or early 2019, consider buying the early 802.11ax APs. These APs will, obviously, be equivalent or better than 802.11ac Wave 2. Please note that these APs will be more expensive than 802.11ac.
- If you already have 802.11ac Wave 1 or Wave 2 APs, consider waiting for the second round of 802.11ax APs to arrive – likely by late 2019 or early 2020 – before starting a replacement. By this time handhelds and laptops are likely to start carrying 802.11ax compliant radios and the cost of 802.11ax APs will have likely dropped to current 802.11ac APs prices.
Want to know more about Mojo APs? Register for a 1:1 demo now.