[AusNOG] Group WiFi

[David J. Hughes]

Awesome!  Thanks a heap for passing those on.  Very interesting read  
(and, errr,  watch).

I think for our event we'll be a slightly better spot than these guys  
are  :-

1. We have a much lower density requirement.  We'll be looking at 100  
to 150 people in a room.  Not 500.

2. We'll have a considerable amount of wired access available.  The  
more people we can get off the .11 network the better.


As AusNOG grows into the hugely popular and successful event we know  
it'll become (confident huh? :)  then we'll need to look further into  
doing QoS at the radio level.



Thanks

David
...

On 02/11/2007, at 9:58 PM, Adrian Chadd wrote:

> On Wed, Oct 31, 2007, Stephen Baxter wrote:
>
>> I need to get wireless access for 130+ people for a couple of  
>> days. If
>> you want to know why then please go to http:// 
>> 2207.ausnog.net :-)   Any
>> advice appreciated.
>
> I replied in private to Stephen but I figure I might as well let  
> everyone
> else know.
>
> Tony (Anton) Kapela does a lot of wireless stuff for NANOG, the  
> American
> network operators group conference; he's put up a few presentations  
> here
> and there covering the sorts of things he does to make wireless  
> "work".
>
> His presentation at a recentish nanog on wireless QoS is interesting
> and actually gave measurable results:
>
> http://www.nanog.org/mtg-0610/presenter-pdfs/kapela.pdf
>
> There's a realmedia recording of his presentation but I haven't  
> tried it:
>
> http://www.nanog.org/mtg-0610/lightning.html
>
> The basic idea is realising that once you've sorted out the RF  
> engineering
> side you still have to deal with a best-effort half-duplex  
> broadcast-mostly
> shared infrastructure.
>
> There's a few nifty tips (eg set to drop on max-retry; not  
> deassociate);
> also see how CoS maps to physical transmission timing. The "TCP  
> initial" stuff
> IIRC is to make sure that TCP connections get a chance to succeed when
> there's a whole lot of potentially very large packets clogging the RF
> space.
>
> I think he's covered many other aspects including handling roaming,  
> handling
> busted wireless clients in laptops (vista's is fun!), RF design  
> tips as
> well as the obvious stuff.
>
> I'm trying to get references to some of his other presentations and
> see if I can't put them all up somewhere which can benefit everyone.
>
> (Note: I've deployed QoS on Cisco APs and measured "stuff", but for  
> VoIP
>  and video streaming in an office environment, not for a conference.)
>
>
>
> Adrian
>
> _______________________________________________
> AusNOG mailing list
> AusNOG at ausnog.net
> http://www.ausnog.net/mailman/listinfo/ausnog
>