Re: virtual wireless interfaces / WDS

On 18-Apr-2003 Jacques Caron wrote:
> At 08:42 17/04/2003, David Young wrote:
>>  An AP can bridge to an
>>   ethernet, however, its clients *cannot* behave as bridges. This is not
>>   a constraint of wi(4) or the BSD 802.11 framework, but a constraint
>>   that the 802.11 committee gave to us. Maybe you are talking about
>>   something else, though.
> 
> Er, why is that? That's the whole purpose of WDS, isn't it? I know it's not 
> normalized by 802.11, but most of the elements are in the spec (the 
> 4-address headers), and with that everything should work well, and I 
> believe the guys using the Linux hostap driver do that every day, don't they?

The key point to remember about a WDS is that there is one bridge between the
WDS and the rest of the world. The fact that packets between the APs in the WDS
may travel over the same physical wire as other ether packets doesn't change
this. If this wasn;t the case how would machines external to the WDS know
how to route to a particular STA?
 
> The only stuff that is really undefined by the spec is how the network 
> topology is learned by the different nodes. This has implications on 
> broadcast traffic which is probably something you want to avoid a lot more 
> on a wireless network than on a wired one, but other than that...

This is defined in 802.11f.

> But maybe we are indeed not talking about the same thing...

D

-- 
________________________________________________________________________
Duncan Barclay  | 
[email protected]   | 
[email protected]| 
--
*bsd wireless list, a bawug thing <http://www.bawug.org/>
[un]subscribe: http://lists.bawug.org/mailman/listinfo/bsd-wireless/