[ovs-discuss] [PATCH] in-band.c: fix for ARP requests and replies for in-band mode

[Nicholas Bastin]

On Thu, Jul 18, 2013 at 6:08 AM, <andreas at brinner.de> wrote:

> > One thing that I think should be pointed out is that it is NOT the
> purpose
> > of OVS in-band rules to forward traffic *for other devices* - the rules
> > exist *only* to ensure that OVS's own control connection functions
> > properly.
>
> Under that premise my patch certainly doesn't make sense. Once the OVS is
> connected to the controller, the controller certainly can add rules which
> enables cascaded switches to connect.
>
> But then the DESIGN paper is at least unclear or misleading
>

As I originally replied, I find the DESIGN doc a little inconsistent about
the goals of the hidden rules for in-band control.  The ARP rules make
sense and losing that level of control isn't unreasonable, as there's no
persistent connection here.  However, the TCP rules hijack the control
connection of other devices (unless you work around this by putting them on
other trasport-layer ports, but that's a configuration nightmare).

I've been looking at this for a couple days and hopefully will have time to
write down some more concrete thoughts in the near term - I've been running
openflow networks with in-band control since certainly at least early 2011
and have experienced a lot of implementations (although we also took a lot
of guidance from the OVS DESIGN document at the time when tweaking things),
and what OVS does at this point seems like over-reaching a bit.

--
Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openvswitch.org/pipermail/ovs-discuss/attachments/20130718/cfe7e6d8/attachment-0001.html>