BROUTING VLANS

BROUTING VLANS

[Asher Awelan]

SCENARIO:
I have inherited a hospitality solution which uses 802.11q tagging on it's switched network for the purpose of room identification, i.e. one port per room.
My current handling of this is working but not stable and is memory heavy.
Currently I have the 250 (I know!) VLANs on eth1.x as interfaces and have bridged them on br0.
I am using ebtables to log which mac addresses are using which vlan tag and then doing some cron processing to charge data use to the rooms.
However, this is pretty intensive on the network/ifconfig management.

QUESTION:
Is there a way to broute the vlans to the eth1 interface and not have to create every interface and register every vlan... or perhaps strip the vlan id (after logging of course)...

Further clarification can be given of course.
I know this is a ridiculous scenario, however these are my parameters within which to work.

Thank you kindly,

Re: BROUTING VLANS

[Grant Taylor]

On 11/15/10 05:02, Asher Awelan wrote:
> QUESTION:
> Is there a way to broute the vlans to the eth1 interface and not have 
> to create every interface and register every vlan... or perhaps strip 
> the vlan id (after logging of course)...

It /might/ be possible to strip the VLAN tags and send the traffic on 
it's way with out the need for all the vlan interfaces.  But, I don't 
know that it will be possible (at least as easily) to re-tag the 
returning traffic.

Something you might consider doing is moving the (un)tagging / trunking 
in to a user space application that can dynamically add and remove the 
VLAN tags with out needing kernel support or vlan interfaces.  -  I'm 
just typing raw theories out my (you know what) and would have to give 
this quite a bit more thought.



Grant. . . .

Re: BROUTING VLANS

[Oskar Berggren]

I'm not sure I realise what you consider to be a problem. Anyway, FWIW
we have several linux routers each with several hundred
vlan-interfaces, I think about 600 at most currently. We are doing ip
routing, but not bridging, for these. Not seeing any problems.

/Oskar


2010/11/15 Grant Taylor <gtaylor@riverviewtech.net>:
> On 11/15/10 05:02, Asher Awelan wrote:
>>
>> QUESTION:
>> Is there a way to broute the vlans to the eth1 interface and not have to
>> create every interface and register every vlan... or perhaps strip the vlan
>> id (after logging of course)...
>
> It /might/ be possible to strip the VLAN tags and send the traffic on it's
> way with out the need for all the vlan interfaces.  But, I don't know that
> it will be possible (at least as easily) to re-tag the returning traffic.
>
> Something you might consider doing is moving the (un)tagging / trunking in
> to a user space application that can dynamically add and remove the VLAN
> tags with out needing kernel support or vlan interfaces.  -  I'm just typing
> raw theories out my (you know what) and would have to give this quite a bit
> more thought.
>
>
>
> Grant. . . .
> --
> To unsubscribe from this list: send the line "unsubscribe netfilter" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>