The minimum basic rate changes the airtime cost of every SSID on every AP, yet it is often chosen by habit rather than by design. Here is how I choose between 6, 12 and 24 Mbps, and why I leave 9 and 18 Mbps alone.
I get asked this question on a lot of projects. Sometimes the network still runs the vendor default. Sometimes it runs a different value in every area, because each change fixed a local problem and nobody went back to align them.
The answer depends on two things: what the minimum basic rate actually controls, and what your AP density can support.
Let us start with the first one, because it is where most of the confusion comes from.
What the Minimum Basic Rate Actually Controls
Three terms get used interchangeably, and they are not the same thing:
- Supported rates are the legacy rates enabled on the BSS. A rate you disable is not advertised and is not used to transmit within the BSS.
- Basic rates are the subset of supported rates that every client must support to join the BSS. They are flagged in the Supported Rates element of the beacon. Cisco calls them mandatory rates; most other vendors call them basic rates.
- The minimum basic rate (MBR) is simply the lowest basic rate.
The basic rates matter because they govern the frames that every client in the cell has to understand:
- Beacons are sent at a basic rate, and in practice at the lowest one. This is the frame that decides how far away a client can still see your SSID.
- Probe Responses are, on most platforms, sent at the lowest basic rate as well.
- Broadcast and multicast are sent at a basic rate. Which one depends on the vendor: many platforms use the lowest basic rate, while some use the highest mandatory rate, unless multicast-to-unicast conversion is enabled.
- ACK and CTS frames sent in legacy format go at the highest basic rate that is less than or equal to the rate of the frame they answer. For HT, VHT, HE and EHT frames, the standard maps the modulation and coding to an equivalent legacy rate for this comparison, so a frame at 64-QAM 3/4 or above counts as 54 Mbps.
And here is what it does not control:
- MCS rates. The legacy rate configuration covers the eight OFDM rates, plus the four DSSS/CCK rates in 2.4 GHz. It does not cover HT, VHT, HE or EHT MCS rates. An 802.11ax client at the edge of the cell can still send data at HE MCS 0, which is 8.6 Mbps on a 20 MHz channel with one spatial stream and a 0.8 µs guard interval, even when your minimum basic rate is 24 Mbps. On platforms that expose MCS controls, those are configured separately.
- Roaming. A higher minimum basic rate shrinks the area where clients can decode your beacons and Probe Responses, which helps them find a closer AP. It does not force a roam. The roaming decision still belongs to the client.
What It Costs on the Air
Let us work one example through, so you can check it against your own tool.
A legacy OFDM frame starts with a fixed 20 µs of preamble and SIGNAL field, followed by 4 µs data symbols. What changes with the rate is how many data bits each symbol carries:
- 6 Mbps: 24 bits per symbol
- 9 Mbps: 36 bits per symbol
- 12 Mbps: 48 bits per symbol
- 18 Mbps: 72 bits per symbol
- 24 Mbps: 96 bits per symbol
Let us take a 400-byte beacon, MAC header and FCS included. Beacon size varies a lot with the elements it carries, so treat this as an example, not a reference value. With the 16 service bits and 6 tail bits added, the frame carries 3,222 bits:
- At 6 Mbps: 135 symbols, so 560 µs on the air
- At 12 Mbps: 68 symbols, so 292 µs
- At 24 Mbps: 34 symbols, so 156 µs
Notice that moving from 12 to 24 Mbps does not cut the airtime exactly in half. The data symbols are halved, but the 20 µs preamble is the same at every rate.
With the default beacon interval of 100 TU (102.4 ms), each SSID sends about 9.77 beacons per second, on every radio. For one SSID on one AP, that is 0.55% of the airtime at 6 Mbps, 0.29% at 12 Mbps and 0.15% at 24 Mbps.
Small numbers, until you multiply them. The cost is per channel, not per AP. Let us take four SSIDs heard from five APs on the same channel: twenty beacon streams. That becomes 10.9% of the airtime at 6 Mbps, 5.7% at 12 Mbps and 3.0% at 24 Mbps, before counting the contention time each beacon waits for.
Beacons are not even the largest cost. A broadcast Probe Request can trigger one Probe Response from every BSSID that hears it. In the large open halls of an event venue, I measured Probe Responses at 24.2% of the frames and 6.7% of the airtime on one channel, against 1.7% for Beacons. Every one of those Probe Responses went out at the minimum basic rate.
The 9 and 18 Mbps Question
The OFDM PHY defines eight rates: 6, 9, 12, 18, 24, 36, 48 and 54 Mbps. Only 6, 12 and 24 Mbps are mandatory. Every other rate, including 9 and 18 Mbps, is optional in the standard.
There is also a technical difference in how OFDM builds these rates. Every OFDM rate is a pair: a modulation and a coding rate. The three mandatory rates all use the convolutional code at its native rate of 1/2: BPSK at 6 Mbps, QPSK at 12 Mbps and 16-QAM at 24 Mbps. 9 and 18 Mbps keep the modulation of the rate below them and reach a higher rate by puncturing the same code to 3/4, meaning part of the coded bits are simply not transmitted.
That redundancy is what makes OFDM robust. The interleaver spreads the coded bits across the 48 data subcarriers, so when some subcarriers fade, the decoder can rebuild the missing bits from the others. A 3/4 code carries less redundancy to do that work. In other words, 9 Mbps is 6 Mbps with less protection, and 18 Mbps is 12 Mbps with less protection.
Nothing breaks if you configure 9 or 18 Mbps as your minimum basic rate, and in practice support for the optional rates is near universal. So the real question is whether they buy you anything.
The standard's minimum receiver sensitivity values show the trade-off: −82 dBm at 6 Mbps, −81 dBm at 9 Mbps, −79 dBm at 12 Mbps, −77 dBm at 18 Mbps and −74 dBm at 24 Mbps. Real radios do better than these minimums, so use them to compare rates, not to predict range.
- 9 Mbps sits 2 dB below 12 Mbps on that scale, and the example beacon takes 380 µs instead of 292 µs.
- 18 Mbps sits 3 dB below 24 Mbps, and the example beacon takes 200 µs instead of 156 µs.
Here is why I prefer not to use them:
- They use a 3/4 coding rate. With less redundancy to recover bits lost on faded subcarriers, a frame at 9 or 18 Mbps is a little harder for clients to decode than a frame at the mandatory rate using the same modulation and a 1/2 code. For the frames every client in the cell has to receive, I would rather keep that protection.
- The standard falls back to the mandatory rates anyway. When no basic rate is less than or equal to the rate of the frame being answered, the ACK or CTS goes out at the highest mandatory rate that is. Building your basic rates from 6, 12 and 24 Mbps keeps your configuration aligned with the rates the standard itself relies on.
How I Choose the Minimum Basic Rate
- 24 Mbps in high density. Lecture halls, arenas, conference centres and dense open spaces. The APs are close enough that a client is never far from one, and overhead is the problem worth solving.
- 12 Mbps everywhere else. This is my default for standard enterprise coverage.
- 6 Mbps only with a reason. Coverage-only designs, where the edge of the cell genuinely needs the extra reach, and legacy clients that do not behave well with a higher minimum basic rate. If you keep it, know that it is the most expensive option per frame.
Here are a few other tips I can give you around basic rates:
- Check 2.4 GHz separately. Disabling 1, 2, 5.5 and 11 Mbps removes 802.11b-only clients from the SSID. Confirm none of your devices depend on those rates before you do it.
- Disable the rates below the minimum basic rate. A lower rate left supported but not basic is still available for legacy data frames.
- Mark 24 Mbps as basic alongside 12 Mbps. With 12 Mbps as the only basic rate, ACK and CTS frames answering frames sent at 24 Mbps or faster go out at 12 Mbps. With 24 Mbps also basic, they go out at 24 Mbps. Because 24 Mbps is a mandatory rate, no OFDM client is excluded. On Cisco, multicast follows the highest mandatory rate, so validate multicast-dependent applications at the cell edge.
- Verify it in a capture. Check the Supported Rates element of the beacon, where Wireshark marks basic rates with (B), then check the rate of Beacons, Probe Responses and ACKs in the radiotap header.
One caveat. The minimum basic rate is not a fix for missing APs. Raise it on a design that does not have the density, and you create areas where clients can no longer decode beacons from any AP. Set the density first, then the minimum basic rate, then validate on site.
Resources
If you want to go deeper on where the OFDM rates come from:
- 802.11 OFDM Data Rates – The Math Behind The Numbers by Renzo Notter. Renzo builds every OFDM data rate from a handful of variables: data subcarriers, modulation, coding rate, spatial streams and symbol time with the guard interval. It is the quickest way to see that 9 and 18 Mbps share their modulation with 6 and 12 Mbps and differ only in coding, and he extends the same math to 802.11n.
- Wi-Fi Probing Behavior and Configured Data Rates by Andrew von Nagy. Andrew walks through captures showing that APs answer Probe Requests at the lowest basic rate, never at a disabled rate, and explains why raising that rate shrinks the association range of an AP but not how far its signal travels. It is a good companion to the Probe Response and roaming points in this article.
Where do you set your minimum basic rate today, and have you found an environment where 6 Mbps was still the right call?