Before you read. A job advert asks for experience with Wi-Fi 6E. A quote from a supplier lists an AX3000 access point. A colleague says the site is “still on ac”.
Three ways of naming the same axis, and none of them appears anywhere in the exam objectives. What is each one saying?
The Network+ objectives name 802.11 as a category and never enumerate a single letter, with one exception: 802.11h, called out for its regulatory behaviour. No generation number appears either. So this track, which follows the objectives, has taken you through five wireless topics without ever telling you what Wi-Fi 6 means, and every datasheet, job advert and supplier quote you meet will assume you know.
Some words you will need
- amendment
- A document that changes the 802.11 standard. The letters after 802.11 name amendments, not products.
- generation
- A Wi-Fi Alliance marketing name, of which only a handful exist. It maps to an amendment.
- spatial stream
- One independent signal path between transmitter and receiver, carried by its own antenna chain.
- MIMO
- Multiple in, multiple out. Several spatial streams at once between the same pair of devices.
- OFDMA
- Splitting one channel into subchannels so several clients can be served in a single transmission.
- multi-link operation
- A client using two bands at the same time, rather than choosing one and associating to it.
- airtime
- The time a transmission occupies the channel. The scarce resource, more than bandwidth is.
What breaks without this
You cannot read a quote. A supplier proposes hardware by generation and by throughput figure, and both are marketing names for engineering that has to be looked up before the price means anything.
You buy a generation your clients cannot use. The access point is one device. The value is decided by the several hundred devices that connect to it, and half of them are three years old.
You cannot tell an upgrade from a refresh. Some steps changed what one client could achieve, and some changed what happens when two hundred of them are in a room. Those are different purchases and they solve different complaints.
Two naming schemes, and only one of them is in the standard
The IEEE publishes amendments to 802.11 and gives each one letters. Those letters are the real identifiers: they name a document, they have a publication date, and what they contain can be read.
The Wi-Fi Alliance is a separate organisation which certifies products, and in October 2018 it announced generation numbers as a plainer way to talk about the same thing. It named exactly three: Wi-Fi 6 for 802.11ax, Wi-Fi 5 for 802.11ac, Wi-Fi 4 for 802.11n.
Two consequences worth carrying. The names below four were never assigned, so anybody saying Wi-Fi 3 has invented it, and the older amendments are still called by their letters because that is all they ever had. And the certification programme runs ahead of the document: the Alliance opened Wi-Fi CERTIFIED 7 in January 2024, and the IEEE published 802.11be in July 2025, so devices were sold and certified against a draft for a year and a half. That is normal in this industry and it is worth knowing before somebody tells you a product predates its own standard as though it were a scandal.
If you already read standards: why an amendment is not a document you can read on its own
An amendment is a set of edits. 802.11ax does not describe how Wi-Fi works; it describes the changes to make to 802.11 that add high efficiency operation, in the form of clause numbers, replacement paragraphs and new subclauses. Reading one on its own is like reading a patch file without the source.
Every few years the working group rolls the accumulated amendments into a new revision, and that revision is the readable document. 802.11-2020 folded in the amendments up to that point, 802.11-2024 folded in more, and each supersedes the last. So the useful reference is almost always the current revision rather than the amendment whose letter people quote.
There is a practical consequence for anybody who wants to check something rather than take a vendor’s word for it. IEEE 802 standards are available at no cost through the IEEE GET program, which asks you to accept its terms and then hands over the PDF. That makes 802.11 one of the very few standards in this whole track that you can read in full without paying, which is worth doing once, if only to see how much of it is not about radio at all.
What actually changed at each step
The generations are not four steps of the same kind. Two of them made one client faster and two of them made a crowded room work, which is a different purchase solving a different complaint.
| Generation | Amendment | What it added | Whose problem it solved |
|---|---|---|---|
| Wi-Fi 4 | 802.11n | Several spatial streams at once, and channel bonding | One client, given clear air |
| Wi-Fi 5 | 802.11ac | Wider channels, denser modulation, 5 GHz only | One client, given clear air |
| Wi-Fi 6 | 802.11ax | Subdividing a channel between clients, scheduling, network colouring | Many clients in one room |
| Wi-Fi 7 | 802.11be | Wider channels again, denser modulation again, and using two bands at once | Both, and latency specifically |
The change at Wi-Fi 6 is the one worth understanding, because it is the one that breaks the pattern. Everything before it made a single conversation carry more. 802.11ax went after the overhead instead: a channel can be divided so that several small clients are served in one transmission rather than queueing for their own turn, overlapping networks can be marked so a radio does not fall silent every time it hears a neighbour, and battery devices can be told when to wake up so they stop waking to hear nothing.
None of that shows up as a bigger number on a speed test with one laptop in an empty office, which is exactly how it gets tested and exactly why the upgrade is often reported as making no difference. Put two hundred devices in a lecture theatre and it is a different machine.
If you already run dense wireless: what subdividing a channel actually buys, and why the gain is largest for the smallest packets
Before 802.11ax, a transmission occupied the whole channel for its duration whatever it was carrying. A client with a 200 byte acknowledgement to send took its turn, used an entire 80 megahertz channel for the time it took, and handed it back. The overhead of taking a turn is fixed, so the smaller the payload the worse the ratio, and a room full of phones sending small things spends most of the airtime on the ceremony rather than the content.
Subdividing the channel lets the access point serve several of those in one transmission, each on its own slice. The cost of taking a turn is paid once and divided.
Which tells you where the benefit is and is not. Bulk transfers were already efficient, because a large frame amortises the overhead by itself, so a file copy sees almost nothing. Chatty, small, latency-sensitive traffic sees a great deal: the acknowledgements, the keepalives, the sensor readings, the interactive requests. In a warehouse full of handheld scanners the difference is not subtle, and in an office where four people are copying files it is invisible.
That is also why the honest way to size the improvement is to count devices rather than to measure throughput. The mechanism scales with how many clients are competing, and any test with one client is measuring the case it was not built for.
The number on the box is a sum
An access point advertised as AX3000 has two radios. Neither of them does 3000 of anything.
The arithmetic behind 2402 is worth doing once, because it demystifies every figure of this kind. A 160 megahertz channel in 802.11ax carries 1960 data subcarriers. The densest modulation puts 10 bits on each subcarrier, of which five sixths are payload after coding, so one symbol carries 1960 by 10 by five sixths, which is 16,333 bits. Symbols take 12.8 microseconds plus a 0.8 microsecond guard interval. Divide and you get about 1201 megabits per second per spatial stream, and two streams is 2402.
Every assumption in that sentence is optimistic. The densest modulation needs a signal strong enough that the receiver can tell 1024 constellation points apart, which in practice means the same room. The widest channel needs 160 megahertz to be free of radar and of neighbours. The guard interval is the shortest one, which needs low multipath. Nothing in it accounts for the time spent not transmitting, which on a busy network is most of it.
What 6 GHz changed, and what it cost
In April 2020 the FCC opened 1200 megahertz in the 6 gigahertz band for unlicensed use, and other regulators followed with their own decisions and their own amounts. That is more new spectrum than Wi-Fi had ever received at once, and it arrived with a property that mattered more than the width: nothing old was on it.
That is the real gain, and it is easy to miss while looking at the megahertz. A 2.4 gigahertz channel has to remain usable by equipment designed in 1999, so every transmission carries compatibility overhead for devices that may not be present. The 6 gigahertz band has no such history, so a network there can assume every device on it is modern.
The costs are physical and regulatory, and topic 30 has the underlying reasons. Higher frequencies attenuate faster over distance and lose more crossing a wall, so the same access point covers less. Regulators restrict indoor power and, for higher power outdoors, require coordination with the licensed users who were there first. And availability is a national question rather than a technical one, so a design that works in one country may have half the spectrum in another.
If you already plan spectrum: why the 6 gigahertz band came with a coordination requirement, and what that means for outdoor coverage
The band was not empty when it was opened. Fixed point to point microwave links, which carry things like utility control traffic and broadcast backhaul, have licensed use of parts of it and were there first. Opening it to unlicensed devices meant accepting a shared band and finding a way for the newcomers not to interfere with the incumbents.
The answer was to split the permissions by risk. Low power indoor operation is allowed without asking anybody, on the reasoning that a building attenuates enough to protect a link outside it, and the FCC’s order defines indoor with a set of physical requirements rather than with a promise: the device cannot be weather resistant, must have an integrated antenna it is not possible to swap, and cannot run on battery power. Read that list again and notice what it is doing. It is a regulator writing a definition a manufacturer cannot argue its way around, because every clause removes a way of using the thing outdoors.
Higher power, which the order calls standard power, is a different arrangement. A standard power access point has to ask an automated frequency coordination system, an AFC, which frequencies it may use at its location. The AFC knows where the licensed microwave links are and returns the channels that will not interfere with them.
For a network designer the consequence is concrete rather than legal. The 6 gigahertz coverage you can rely on indoors is smaller than the 5 gigahertz coverage it sits alongside, by more than the frequency difference alone would suggest, because the power ceiling is lower too. Outdoor 6 gigahertz coverage is a different class of deployment with a dependency on a service that has to be available in your country. Neither of those is a reason to avoid the band. Both are reasons the channel plan and the mounting positions are a fresh piece of work rather than an inheritance.
The generation that counts is the one your clients have
An access point is one device. A network is that device and every laptop, phone, badge reader, printer and handheld scanner that associates with it, and the distribution of what those support is the thing that decides what the network can do.
There is a physical reason this matters more than it sounds. The scarce resource is airtime, not bandwidth. A client transmitting at a low rate takes longer to send the same frame, and while it is transmitting nobody else can. One old device at the far end of a warehouse, sending slowly because it is far away and because it is old, can consume more channel time than a dozen modern clients close by.
So the question to ask about any wireless upgrade is not what the access point supports. It is what fraction of the devices that will connect to it can use the new thing, and whether the ones that cannot are going to be replaced or are going to be there for another five years holding the channel.
Prove it
Do the rate arithmetic yourself. 980 data subcarriers on an 80 megahertz channel, 10 bits each, five sixths of them payload, in a symbol lasting 13.6 microseconds including the guard interval. Work it out and you should land within a megabit of 600, which is the figure a single-stream client on a good 80 megahertz connection reaches. Then find any access point datasheet and check that its headline number is the sum of its per band figures.
Read the announcement rather than an article about it. The Wi-Fi Alliance’s own release from October 2018 is short. Read it and count how many generation names it defines. The answer explains why nobody can tell you what Wi-Fi 3 was.
Look up an amendment’s status. The IEEE 802.11 working group publishes a timelines page giving every task group, its current draft, and its projected completion. Find 802.11bn on it. It is the one that will be called Wi-Fi 8, and the projected approval date tells you how much of what you read about it today is a prediction.
What trips people up
1. Treating the generation as a speed
It is a set of capabilities with a date on it. Two Wi-Fi 6 access points can differ by a factor of four in throughput depending on how many antennas each has and which bands it covers.
2. Reading the box number as achievable
It is the sum of every radio’s theoretical maximum. No client uses more than one radio at a time before Wi-Fi 7, and no client reaches a theoretical maximum outside a shielded room.
3. Thinking 6E is a new protocol
It is Wi-Fi 6 with a 6 gigahertz radio. The engineering is 802.11ax either way, and the difference is which spectrum the device is permitted and able to use.
4. Assuming the newest generation is available everywhere
Spectrum is a national decision. A 6 gigahertz design validated in one country can arrive at a site in another with substantially less spectrum to work with, and the channel plan has to be redone rather than copied.
5. Upgrading the access points and not the survey
A newer generation at 6 gigahertz covers less ground per access point than the 5 gigahertz network it replaced. Reusing the old mounting positions produces coverage holes that the old hardware did not have, which reads as a fault in the new equipment.
6. Testing the upgrade with one laptop
The gains from Wi-Fi 6 onward are mostly about serving many clients efficiently. A speed test on an empty network measures the thing that did not change much.
Work it through
A school replaces its access points with Wi-Fi 6E hardware, mounted where the old ones were. Complaints go up. The IT lead reports that a speed test in the corridor gives the same number as before, and that some rooms are now worse.
Take the two symptoms separately, because they have different causes.
The corridor speed test being unchanged is expected and is not evidence of anything. One device on a quiet network was never the case 802.11ax improves. The measurement that would show a difference is thirty devices in one classroom during a lesson, and the number to read is not throughput but how long a page takes to start loading.
The rooms that got worse are the interesting half. If the new radios are serving 6 gigahertz, they cover less than the 5 gigahertz radios they replaced, from the same mounting points, through the same walls. That is physics rather than a fault, and it means the estate needed a new survey rather than a like-for-like swap.
The check that settles it is to look at which band the complaining clients are actually on and what signal they report. If the bad rooms are the far ones, and the clients there are hanging on to a weak 6 gigahertz signal rather than falling back, the fix is coverage: more access points, or a policy that steers distant clients down to 5 gigahertz where the reach is longer.
Then the question nobody asked at purchase. How many of the school’s devices can use 6 gigahertz at all? If it is the newest third of them, the other two thirds are on the same 5 gigahertz radios as before, sharing the same air, and the money bought a better network for the tablets bought last year.
Try it
Find out what your own machine supports. On Linux, iw dev names the
interface and iw dev <name> link prints the band, the channel width and the
negotiated rate. On macOS, hold Option and click the Wi-Fi icon for the mode,
channel and rate of the current connection. On Windows,
netsh wlan show interfaces prints the radio type and the receive and transmit
rates. All three will tell you which generation you are actually using, which is
often older than the one on the router.
Read the rate your client negotiated, then walk away from the access point. Watch it drop as the modulation gets less dense. That single observation connects the arithmetic above to a physical distance, and it is the fastest cure for treating the headline figure as a property of the equipment.
Look at a datasheet and take the number apart. Any access point, any vendor. Find the per band figures, add them, and confirm they make the name.
Check yourself
Somebody says a site is running Wi-Fi 3. What have they got wrong?
There is no Wi-Fi 3. The Wi-Fi Alliance defined three generation names in 2018, for 802.11n, 802.11ac and 802.11ax, and numbered them 4, 5 and 6. Everything older is referred to by its amendment letter because it never received a number.
What is the difference between a Wi-Fi 6 and a Wi-Fi 6E access point?
A radio for the 6 gigahertz band. The protocol is 802.11ax in both cases. 6E is a Wi-Fi Alliance name for Wi-Fi 6 equipment that can also operate in the spectrum regulators opened from 2020, not a separate amendment.
An AX3000 access point is installed and a laptop reports 1200 megabits. Is something wrong?
No. 3000 is the sum of the two radios’ theoretical maxima. The laptop is on one band, and 1200 is close to what two spatial streams on a 160 megahertz channel reach, so it is doing well rather than badly.
Why can one distant old client hurt everyone else on the same access point?
Airtime. A client transmitting at a low rate occupies the channel for longer to send the same frame, and nobody else can transmit while it does. Its cost to the network is measured in time rather than in the bandwidth it consumed.
Why do the gains from Wi-Fi 6 rarely show up in a speed test?
Because a speed test is one client on a quiet network, and that is the case 802.11ax changed least. Its mechanisms reduce the overhead of serving many clients at once, so the difference appears in a full room and not in an empty one.
References
- Wi-Fi Alliance introduces Wi-Fi 6 - Wi-Fi Alliance, October 2018, the announcement that defines Wi-Fi 4, 5 and 6 and no others. Free. Accessed 2026-08-20.
- Wi-Fi Alliance brings Wi-Fi 6 into 6 GHz - Wi-Fi Alliance, January 2020, for what 6E names. Free. Accessed 2026-08-20.
- Wi-Fi Alliance introduces Wi-Fi CERTIFIED 7 - Wi-Fi Alliance, January 2024, the certification programme that opened before the standard was published. Free. Accessed 2026-08-20.
- IEEE 802.11 Timelines - IEEE 802.11 Working Group, publication dates for each amendment and the projected dates for the ones still in draft. Free. Accessed 2026-08-20.
- IEEE 802.11 - IEEE Standards Association, the standard itself, for the physical layer parameters the rate arithmetic uses. Accessed 2026-08-20.
- IEEE GET Program - IEEE Standards Association, how to obtain an 802 standard at no cost. Free. Accessed 2026-08-20.
- FCC Report and Order 20-51 - Federal Communications Commission, April 2020, the decision that opened 1200 megahertz at 6 gigahertz for unlicensed use in the United States. Free. Accessed 2026-08-20.
Where the figures came from. Nothing on this page is captured. The generation names and their dates are from the Wi-Fi Alliance’s own announcements, the amendment publication dates are from the IEEE 802.11 working group’s timelines page, and the 6 gigahertz allocation is from the FCC’s order. The throughput arithmetic is worked from the physical layer parameters in 802.11 and can be checked against any access point datasheet, which is the exercise in Prove it. No wireless output in this track is captured, for the reason topic 29 gives: the lab behind these notes has namespaces and no radios.
Why this is not in the lesson count. The objectives name no 802.11 letter except 802.11h and no generation number at all, so none of this is examinable. It is here because every datasheet, quote and job advert you meet outside the exam uses these names and assumes you know them.