Adjusting firmware to automatically switch between wifi standards

Dear all,

I have read through the overview, hello world, Build system and Building OpenWrt for Netgear WNDR3700 guides.

The developer page contains so much information with acronyms I do not yet know. Next to this OpenWRT is so versatile that I am unsure what applies to my specific case.

I would like to know if it is possible to adjust OpenWRT firmware of a given router to automatically switch the wifi standard (b/a/g/n/ac) that is used when the required speed increases.

I am looking into this for my bachelor.

I am familiar with programming in a couple of languages, but my bachelor has been more of the theoretical side of computing science than actual large projects.

If you have any questions, please feel free to ask :slight_smile:

Thank you for reading.

The physical data rates are variable, according to the modulation used, and each subsequent standard is compatible with all previous ones (with some minor caveats). Maybe you should study a bit more about 802.11 (this would be a good place to start). :wink:

1 Like

I know that they are backwards compatible however they will be limited to the speed of the client/STA if the router has ac a client with g can connect but only at g speeds.

I would just like to know if it is possible to limit the standard used and possibly automatically adjust the standard used based on the speed requested/required by the clients/STAs.

I'm not sure I'm following your question. Most AC routers these days (and past 5-7 years) have chipsets that will regulate airtime fairness based on the type of clients connecting, or if the clients support it's it, MU-MIMO.

It's also a big myth that "older devices connecting = everything gets slower". That's not true. Your router adjusts speeds based on the rates negotiated by clients, and to that client ONLY. Modern clients that can connect at their fastest, will connect at that rate. The only thing that could potentially slow the network (on that channel) is the amount of airtime older clients need to "talk" to the router, which can eat up CPU time. Again, very minimal.

1 Like

Thanks for your reply. I know that an older device connecting wont slow down other devices. I meant that only the connection of a client using g wifi standard will be at g speeds and not ac.

I am rounding up my bachelor degree and need to write a small thesis. I like the subject of green ICT. So together with some professors i discussed looking at Wi Fi and the possibility of controlling which the Wi Fi standard that is being used to save in power. Or apply a feature that is like Power Saving mode in most phones. I know it will be a very small saving. (a possible result is no saving ofc) It is just a very small research.

I would like to alter OpenWRT to use 'a' for example and when it gets busy scale up to n and then ac.

I would like to find out here if it is possible. Because if i start and it ends up that its not possible to force specific wifi standards on a router it would be for nothing.

Would that save energy at all? Lower data rate means more time spent transmitting. Assuming the same transmit power, your suggestion would use more energy.

While i don't really see the merit (sorry, good luck with your thesis), in the 5GHz case you can use HTMODE to control the setting.
NONE = a
HT20/40 =an
VHT20/40/80/160 =anac

If you also use the minimum connection setting (can't remember, its in the wiki), you can effectively remove the backwards compatibility.

This applies in 2.4GHz also.

While that is true, the radio will likely require less power. Next to that it is inline with the approach done in EEE 802.13az where power increases almost linearly with data transferred.

So the goal is not to keep the slow network when you want speed. But maintain the slow when slow is the only thing that is needed.

While i don't really see the merit

Do you mean you dont see the added value or benefits?
If so, i get that. If you have any pointers/tips or just your opinion I would love to hear it.

Thank you for your reply :slight_smile:
I will take a look at HTMODE

You might also be interested in playing with the transmission power, for example based on the signal level received by the router.

Both, but i also don't think it has enough substance for a bachelors thesis. It kind of feels like you could bang it out in an afternoon. It would be more interesting to do self organising meshes.

My opinion of course, and im not your supervisor/lecturer. Nor did i do CompSci.

Theres also the issue of having to teardown the wifi network to change its mode. I dont think you can change it on the fly, so this causes all devices to need to reassociate.

Actually, that reminds me… does anyone have any idea about the status of Minstrel-Blues? Now, that's really neat (on paper, at least).

1 Like

Your input is appreciated! It is only a 12 EC thesis. Pretty small, changing just OpenWRT is only the last bit. Before comes a literature review, measurements and other ways to tackle the problem.

1 Like

Thank you, i will take a look :slight_smile: