[Bridge] Re: [ethernet bridge] 802.1G ampliation

[Bridge] Re: [ethernet bridge] 802.1G ampliation

[Stephen Hemminger]

On Tue, 28 Oct 2003 12:24:16 +0100
José J. González Alonso <jjgonzalez@igalia.com> wrote:

> Hi! 
> 
> I am very interested in your project. Concretely, I am interested in the
> possibility of extending your implementation so that it fulfills the
> specifications of 802.1G. I read this specifications and they seem
> "simple" extensions of 802.1D standard.
> 
> what do you think about? 
> would be it very hard?
> Is there any documentation of your code? 
> (to read the code directly is being a little tedious)
> 
> Regards,
> 	José J. González Alonso
> 

There have been several discussions about this, but very little in concrete work.
The bridge mailing list might have others that have done something.

A long term goal is to move all the discovery and spanning tree out of the kernel
and into a user space process.  This user space process would use netlink to
see interface state changes and as API to the bridge driver.  Also a new implementation
in user space could use the standard LLC interface to read bridge spanning tree traffic.

Unfortunately, I don't have a lot of time to dedicate to new feature work right now,
especially when 2.6 is trying to get released.

[Bridge] Bridge that do MSS CLAMP

[David GLAUDE]

In order to transparently help packet go inside a vpn/tunnel without 
having to solve the security problem of permiting ICMP packet getting 
accross a firewal...

I would like to use "clamp-mss-to-pmtu" inside a bridge (firewall bridge).

It would bridge every traffic, but alter the Maximum Segment of TCP 
connection.

Before attempting this, could anybody give us a hint on the possibility 
for this to work... Was this attempted before (or should I ask for a 
"software patent" on this) :-)

It sound a little bit like what a "Packet Shaper"TM does by altering 
windows size computation.

David GLAUDE