Preventing WDS links / Re: WDS configuration issues

[Dave]

Thanks Mike...that makes sense!

I would assume that I would definitely want to use Repeater mode if that is
indeed what it does....bring up a WDS link without allowing clients to
associate...that would pretty much force them to the omni and I wouldn't
have to worry about them accidently hitting the backhaul link.

Then I could continue to use layer 2 bridging as I had planned and bridge
the WDS/Repeater/Backhaul radio iface with the Master radio iface. Makes
sense...and I will give it a shot.

If anyone else can confirm/deny Repeater operation to make sure I am on the
right track it would be greatly appreciated.


Thanks!

-Dave


> On Wed, 12 Feb 2003, Dave wrote:
>
> > Does anyone have an idea on how to PREVENT a WDS link from coming up or
> > associating?
> > Specifically I have a WDS box with two radios, one attached to a yagi
> > pointing to next closest WDS box and the other connected to an omni.
>
> I think you can add the omni's radio to the bridge, but not enable WDS on
> that card. Would that do it? I'm assuming WDS links need to be symmetric
> here, but I could be wrong.
>
> Re: your next message: If you set the backhaul (yagi) link to STA mode
> instead of Master, you won't be able to use WDS or bridging, and you'll
> have to do routing. Not a big deal though - set up zebra's RIP routing
> daemon, figure out some method for assigning unique subnets and ESSIDs
> (static config should work for smallish nets), and the routing will take
> care of itself.
>
> With Master mode on and other_ap_policy=1, you should get the same link
> quality stats you would in STA mode. The numbers/units will be different
> but they'll be there. I've never tried Repeater but I think it allows WDS
> without allowing clients to connect, which is likely what you want.
>
> The channel spread you'd want to use might be more like 1-6-11 than
> 1-8-11, since the 8 & 11 are pretty close. Even with 1-6-11 you'll get a
> little bit of interference between the radios in close proximity (i.e. in
> the same box), but not much.
>
> Mike
>
> _______________________________________________
> HostAP mailing list
> HostAP at shmoo.com
> http://lists.shmoo.com/mailman/listinfo/hostap
>