APs and multiple BSSID support (OT?)
Mon Feb 12 14:55:53 PST 2007
> -----Original Message-----
> Jouni Malinen
> > share many attributes such as timing, channel, rates, etc.
> There is no guarantee that Beacons from the same physical radio will
> share these attributes.. As far as channel is concerned, it will most
> likely be the same and also TSF is likely to be shared. However,
> advertised rate sets could be different..
> Which timing problems are you worried about? If I understood correctly,
> you did not require power save support which makes this much easier..
> Even WMM could be mostly doable assuming the APs are using more or less
> compatible settings (or the client radio supports large enough set of TX
I pretty much agree with everything you said. My worry is pretty general.
The entire .11 protocol depends on exquisite timing and I worry that there
are subtleties that I'm unaware of. I prefer not to take the risk of
committing to the completely general implementation. As it is, I have had
to make changes that involve hundreds of references to data structures
whose usage pattern I am changing. Having two completely unrelated virtual
clients would require much more extensive changes. Handling different WMM
settings is far beyond what I can do in a reasonable amount of time.
Another engineer is doing the UI for the project, but I'm doing everything
> I would expect that this type of virtual client could associate with
> multiple BSSes from multiple APs as long as these BSSes are using the
> same channel. Whether everything works exactly in the same way as it
> would with multiple radios (one per client STA) is another question and
> how important this is depends on the use.
> I don't think there is generic mechanism for determining whether there
> is single physical radio or not.. As an example, I'm aware of multiple
> different ways of doing this with Atheros hardware and I know that more
> than one mechanism has been deployed.
> > If there is another way to determine whether 2 beacons or probe
> > responses with different BSSIDs come from the same physical AP, I
> > would love to hear of it.
> This would get complex if all cases would need to be caught without
> co-operation from the APs. I'm not sure whether this is worth the
> problem taken into account the virtual clients are likely to work well
> enough with multiple APs for most uses.
Again, I pretty much agree with you. That's why I asked the question. If I
know of many specific and incompatible methods, I may have to allow
connection to 2 apparently unrelated SSIDs.
I know exactly what the Atheros HW can do regarding BSSIDs. I know that any
of the 47 bits (other than the broadcast/multicast bit) in the address
might be used in the multiple bssids for a set of virtual APs.
You say you are aware of multiple ways BSSIDs have been deployed with
Atheros hardware. Could you please tell me the vendors and the methods each
used? This information would help. I have only seen the two methods I have
More information about the Hostap