Channel Widths, RUs, and Alphabet Soup
Warning
This blog comes with a warning, it’s suitable for bedtime reading as it might have you off to sleep quicker than counting sheep. This is about 802.11ax channel widths, RUs (resource units) MU-MIMO (multiple user multiple input multiple output) MU-PPDU (multi-user physical protocol data unit) and other alphabet soup that goes along with it. If you’re a non-technical reader stick with this one, it’s easier to understand than on first glance.
History around Wi-Fi Channel Widths, Antennas & Spatial Streams
Up until 802.11ax each STA (client) transmission took up the entire channel width allocation which is known as OFDM (Orthogonal Frequency Division Multiplexing). This was fine at the time of the mid-2010s before we had the IoT boom, AR, VR, video and voice calling caused by the coronavirus and other services which required much more efficient traffic delivery.
Next is important to understand with how antennas and spatial streams are explained to us which for this I am going to use an example 4x4:4:
• The first two numbers are the transmit and receive radio chains which are connected to antennas, these are tied to the 2.4GHz, 5GHz and 6GHz radios (if dual-band and tri-band radios are in use) themselves.
• The last number separated by the colon is the maximum number of spatial streams supported on the antennas.
To explain this further, let’s say we have a bowl of soup and some spoons, as wireless is half duplex, we can either bring soup to our mouth or return an empty spoon to the bowl. Now what our spatial streams do is let us use more spoons at the same time to eat the soup quicker, in this case the STA (client) can use up to 4 spoons to eat the soup at once, but it is important to remember this is four individual spoons, not a spoon four times the size (the number of spatial streams is dictated by the client under normal circumstances as it is less than the AP in most cases, but this can be negotiated during association based on the STA capability).
Now we have some history and background on wireless transmissions pre 802.11ax/Wi-Fi 6 we can move on to the methods implemented by the IEEE’s 802.11ax amendment.
RUs (Resource Units)
With the development of 802.11ax which introduced OFDMA (Orthogonal Frequency Division Multiple Access) to make transmission more efficient by allowing multiple STAs to send and receive simultaneously at the same time by sharing the available bandwidth of the channel.
RUs (resource units) in 802.11ax divide the channels up into smaller subchannels, resource units which are 2MHz wide, these are further broken down into tones which are 78.125kHz spaced apart; 26 tones make up an approximately 2MHz wide RU, which is significantly smaller than in 802.11ac. The larger the RU, the more tones it has within it.
The idea behind RUs is to make transmission of data more efficient by allocating a smaller “sub-channel” that matches the current STA requirements. The width of the RU is dynamically assigned by the AP and is something administrators cannot control but can influence with QoS; the AP learns the client’s buffer status by using buffer status reports and trigger frames. The RU groups are made up of 26, 52, 106, 242, 484, 996 and 996x2 tones wide depending on channel width, with 484, 996 and 2x996 tone RUs only available with 80MHz and 160MHz wide channels.
The smaller RUs are, under normal circumstances, used for latency-sensitive traffic, whereas wider RUs can be used for higher bandwidth applications, but the AP may choose to use MU-MIMO (multiple user multiple input multiple output) rather than OFDMA or it may use both in combination together depending on STA capabilities and traffic patterns, transmitting to multiple clients across both frequency and spatial streams simultaneously. Again, this is something the AP will dynamically choose rather than be influenced by the administrator.
Normally when asked the question “so each client gets the whole 20MHz channel” I will usually answer not exactly, although factually incorrect and I know this, it helps non-Wi-Fi techs and non-technical people understand the concept better. I always use pies or tarts with bandwidth, everyone gets a slice unless it’s treacle.
RUs & MU-MIMO
A quick intro to MU-MIMO, DL-MU-MIMO (downlink multiple user multiple input multiple output) was first introduced in 802.11ac wave 2. At a very high overview it allows the AP to communicate with up to 8 clients simultaneously, it achieved this by using different spatial streams to communicate with STAs on the same channel band.
This at the time was a game changer in wireless. I personally remember being genuinely excited and a little confused when reading about it, it wasn’t until I started working on it that it clicked.
OFDMA and MU-MIMO are more complementary rather than competing technologies due to how they both work. OFDMA with RUs dividing up the channel, while MU-MIMO uses multiple spatial streams to transmit to clients simultaneously. In 802.11ax both technologies can operate together within the HE-MU PPDU (High Efficiency Multi-User Physical Protocol Data Unit) frame. This to me only reinforces how they complement one another rather than replace each other.
It’s also worth noting that uplink MU-MIMO (UL-MU-MIMO) wasn’t introduced until 802.11ax.
To make this easy to visualise, OFDMA is your local parcel delivery van, lots of parcels for lots of addresses. MU-MIMO is similar to a fleet of HGVs, each HGV has its own dedicated load for a specific destination, like a large warehouse. Both are moving data, just in very different ways.
This is again why I often say “not exactly” when I get the question, “So each client gets the whole 20MHz channel?” Without a frame capture, you can’t be certain which method the AP is using at that moment.
MU Frames
What makes this possible in 802.11ax was the introduction of MU-Frames such as:
· MU-PPDU (multi-user physical protocol data unit)
· MU-RTS (multi-user request to send)
· Multi-STA Block ACK (Multi-STA block acknowledgement introduced in 802.11e but employed by 802.11ax)
· HE-TRG (High Efficiency trigger)
These are the real nuts and bolts of Wi-Fi at this level; they’re what coordinate the whole OFDMA and MU-MIMO orchestra, ensuring everyone is reading from the same hymn sheet.
At this point, we’re deep into the engine of how 802.11ax really works, where it’s all bits, timing and coordination. For the majority of us, it’s enough to know that this is what makes modern wireless highly efficient. The rest is quite literally 1s and 0s, that’s where I enjoy sitting.
To infinity and efficiency
With 802.11ax approaching 6 years old and 802.11bn now becoming more mainstream, the use of the spectrum is inevitably going to become more efficient and better co-ordinated. I am just in the finishing sections of Wi-Fi 7 in depth book which has been a fantastic read during my morning commute, but I am yet to write on the subject due to changing jobs and other personal commitments.
What 802.11ax highlighted to me more is that Wi-Fi sometimes is less about raw speed and more about coordination, timing, precision, balance and communication. Seeing how this can increase efficiency is what’s made me want to implement it into my professional career, but I digress. For me, understanding what is going on under the surface is only half the fun, it’s what turns “just another IEEE standard” into something fascinating, and understanding what some of the 1s and 0s do gives me a huge amount of satisfaction.