• Are there any best practices for 802.11a/n and 802.11b/g/n data rates as far as which to set as mandatory/supported etc?

    I'm looking at optimizing our current deployment and have started by disabling some of the lower data rates in the 2.4 range - everything below 12 Mbps.

    Current setup is as follows:

    1 - Disabled
    2 - Disabled
    5.5 - Disabled
    6 - Disabled
    9 - Disabled
    11 - Disabled
    12 - Supported
    18 - Supported
    24 - Mandatory
    36 - Supported
    48 - Supported
    54 - Supported

    6 - Mandatory
    9 - Supported
    12 - Mandatory
    18 - Supported
    24 - Mandatory
    36 - Supported
    48 - Supported
    54 - Supported

    I'm not worried about b clients - we don't have any - and even if we did I don't want them on the network. :)

  • I would guess it depends on how dense your network coverage is and what speeds your clients support upto. Switch of too many lower rates and clients will be constantly scanning and reassociating to a new AP, kind of defeating what you are trying to do.

    Try doing an active survey when the office(?) is empty and switch off some more data rates and see what happens. If you have a good WNMS it should be easy enough to see how the clints moves about. If clients are static, you may get away with only higher rates, but if your clients are on the move (smartphone, PDA, Voice) it may be more of a head ache!

    Would be interested to know the outcome, including equipment used and facility type.



  • Just remembered, we installed a SCA network a while ago with reasonable density and left 18Mbps upwards on and never had any bad reports.


  • This brings up the subject of what exactly "Mandatory" means with those rates being 6, 12, and 24 Mbps for 802.11g.

    I always thought I knew what that meant until I did some recent testing. It seems some WLAN hardware may get confused and not be able to send ACK's, if all three aren't supported. I had always assumed it meant that manufacturers had to support these on every product, even if the user didn't enable them. But now I'm not as sure.

    The 802.11g standard is a little obtuse in this area. Has anyone out there had an authentic problem with this?

  • We typically disable 1,2 and 6 in our WLAN deployments. We use VoWIFI and find that the lower data rates cause more issues. We prefer the clients to roam to better coverage that stay connected at the lower rates. As far as issues with ACKs, I haven't ran into that but I try keep to the same vendor equipment.

  • It's important to note why a person would be disabling data rates.

    One important thing is that it will shrink the effective cell size but it does not shrink the actual cell. No matter what data rate, the RF still travels under the same amount of power (for the most part).

    So, setting higher data rates does not reduce co-channel interference.


  • Great tips for optimizing your current deployment can be found in one of Marcus' postings: (in this part he also discusses data rates).

    Also on cisco live 2011 there was a session on performance tuning of wlans where this was mentioned (and tweeted/blogged by @revolutionwifi) : (see also the articles mentioned at the bottom of the blog)

    My 2 cents:
    I disabled everything below 12 mbps on both 802.11a/g with rather spectacular results. We do not have control over the clients (education) and many clients connected on way to low rates. One one site i will soon disable everything below 24 Mbps. We don't do vowlan, just data.

    Best regards, RU

  • The IEEE required only mandatory support of data rates of 6 , 12, 24 Mbps, the reason behind this, because each of these rates used a different modulation ( BPSK,QPSK and QAM). 

  • By Howard - edited: January 4

    Maybe that's one reason, but perhaps more significant is that each of those three headline rates is "un-punctured (UP)" - and 24 Mbps is the highest un-punctured rate. 

    All of the headline rates above 24 Mbps are punctured, which is essentially compressing the number of error correcting bits in a "symbol".

    I have noticed, when performing rate-vs-performance tests, that graphs sometimes  "jump around" at the UP rates.  Probably because the UP rates use less robust encoding.

    Look at the "coding rate" parameters for each headline rate and you'll see the difference.

    Matthew Gast's "Definitive Guide" has a good section on puncturing.

Page 1 of 1
  • 1