Same SSID for both 2.4 and 5GHz bands?

[sfx2000]

I was referring more to when you have 2+ access devices, same band.

eh?

Perhaps I'm missing something here, or an incomplete thought...

 

[azazel1024]

eh?

Perhaps I'm missing something here, or an incomplete thought...

I think we may just be missing something.

Its not hard for the client to check the RSSI of the different APs and the different bands (as well as how many radios) and choose which AP and which band to connect in. Its a rather trivial problem for a client to figure it out and they've been doing it for at least a few years now.

I don't have a ton of dual band, multi-AP on the same SSID experience, but I have plenty stress testing my single band multi-AP on the same SSID and my clients have no issues picking which is the faster AP at a given time. My limited testing on multiple APs with dual band and the same SSID, they also didn't have any issues switching bands and choosing the faster AP.

At worst one might hold on to an AP/band a little longer than maybe was ideal, but it never devolved to an untenable situation before it switched.

The only device I have that seems to have an "issue" is my iPhone 5, where it takes a couple of minutes to promote back up to 5GHz once the signal strength is better for that (and on the iPhone 5, its 65Mbps on 2.4GHz and 150Mbps on 5GHz). All my other clients I've tested both switch APs and switch bands within about 7-15s of one band or AP being the better choice. If you are moving really far beyond a band or AP being ideal to another one it'll switch a little faster.

This is with an outdoor AP, an indoor AP and an indoor router all on 2.4GHz for my extensive testing of network connectivity. I've tried iPad 2s, iPhone 4s, iPhone 4, iPhone 5, Asus T100, Intel 1000, 2230 and 7260AC wifi cards (oh, and an Atheros of some breed) and Asus Memo Pad HD 7s and all "roam" well between the APs with very similar behavior.

They'll switch within 15s, generally within 10s of one AP being the better one and it rarely ever interupts anything. I've seen VOIP interupted once out of the two times I've tried it. I've seen a video stream (Youtube, I think) interupted once in the dozens of times I've roamed while streaming youtube or netflix and I don't think I've see any CIFS file transfers catastrophically interupted in the dozens of times I've done it (they do often slow down or "hang" for maybe a second before resuming), and I think I've seen facetime disconnect a couple of times in a dozen or more roaming attempts. I've probably seen a couple of page loads hang while roaming of the dozens to hundreds of times I've roamed between APs in my house/outside of my house.

The 2.4/5GHz testing again was just at my brother's apartment which has a pair of APs at either end (Netgear WNDR3400 I think) and my T100 and iPhone 5 seemed to seamlessly transition between them and between bands, though again, the iPhone 5 hung on to 2.4GHz for a long time before transitioning.

 

[stevech]

I think we may just be missing something.

Its not hard for the client to check the RSSI of the different APs and the different bands (as well as how many radios) and choose which AP and which band to connect in. Its a rather trivial problem for a client to figure it out and they've been doing it for at least a few years now.

Some do, some don't.
For moving client devices (e.g., handhelds) many stay with a weak signal even though there's a friendly AP nearby with a better signal.
Etc., etc. This is the classic issue of the low level firmware not wanting to start scanning channels to find a better access device and risk disrupting layer 3 protocol work.

 

[azazel1024]

Sure it happens, but in practice I haven't seen any really terrible offenders, at least not with iDevices or with the couple of Android tabs I've played with.

They tend to switch roughly at the same locations and roughly as quickly as my windows devices do, where I can easily monitor RSSI.

Generally any device needs an AP signal strength to be at least 10dB higher than the current AP before it'll ever switch. Its a little more tricky as my AP is on 40MHz 2.4GHz and my router is 20MHz 2.4GHz.

They tend to switch from my AP (40MHz) to the router (20MHz) when the router has an RSSI roughly 15dB higher than the AP. The switch back tends to occur when the AP has a RSSI 10dB higher than the AP.

So it isn't perfect, but it is pretty seamless and there aren't any cases where RSSI drops so low that there is a significant drop in wifi speed before it'll switch to the other AP. It just might be a case where it goes from 65Mbps to 150Mbps instead of transitioning before the current AP RSSI drops low enough that the speed is below the about to connect APs speed.

Using seperate SSIDs I've found that the roaming behavior and AP choice is different and poorer. I find that Wifi devices often hang on to the existing AP/router even longer/greater RSSI imbalance than they do in a multi AP, same SSID scenario. I can't speak to band switching, but I'd assume similar behavior.

I am not throwing my outdoor AP in to the switching game because pretty much the moment you step outside and close the door outback the indoor router goes from around -45dB to -60dB and the outdoor AP jumps from -65dB to -40dB, so there isn't an easy test for "see when it'll switch APs" because we are talking a difference of 25dB pretty suddenly so it switches within seconds and with no further change in RSSI needed.