FreeVRRPd project status
Claudio Jeker
cjeker at diehard.n-r-g.com
Wed Apr 13 13:08:25 PDT 2005
On Wed, Apr 13, 2005 at 09:36:48PM +0200, Lars Erik Gullerud wrote:
> On Wed, 13 Apr 2005, Claudio Jeker wrote:
>
> >On Wed, Apr 13, 2005 at 05:14:52PM +0200, Lars Erik Gullerud wrote:
> >>
> >>...and can't safely be deployed in a lot of datacenter scenarios where
> >>the providers gear is running VRRP, since the OpenBSD-folks didn't bother
> >>to read up on how the process of obtaining a protocol number works, and
> >>hence used the one assigned to VRRP after a half-baked attempt at getting
> >>one themselves. Hence making CARP pretty much useless for ISPs, no matter
> >>how good it may or may not be otherwise.
> >>
> >
> >This is not true. First of all the "OpenBSD-folks" asked IANA for protocol
> >numbers for CARP and pfsync but IANA denied it. The reason was that CARP
>
> Which is exactly what I said, they didn't bother to read how the process
> works and accordingly made a half-baked attempt only. You don't just fire
> off a mail to IANA and say "hi, can I get a protocol number", that's just
> not how things work, except in OpenBSD-land, obviously. :)
>
OpenBSD did more than just write a mail "hi, can I get a protocol
number".
> >was not developped through an official standards organization.
>
> Which is balony, you do however need to take the PROCESS through
> the correct "organization" (i.e. the IETF and friends, although the
> protocol itself can be developed through my grandma's knitting club). So,
> I stand by my initial statement (but hey, I'm a network engineer at an
> ISP, not a BSD developer - yes, us people the OpenBSD guys don't like
> much because we like to point out the glaring problems in things like
> CARP and OpenBGPd). However, this is all very much beside the point, so
> further IETF/IANA-bashing or OpenBSD-bashing can be taken somewhere more
> appropriate than this list. (Feel free to flame me privately)
>
The problem is that the correct "organization" is a exclusive club where
opensource mostly doesn't matter. It is hard to get into the club as a
non-profit project that mostly gets developped in spare time.
Also I would be intrested what the glaring problems of OpenBGPd are.
An important part of developement is feedback.
> My point is that this very unwise decision to reuse the VRRP protocol
> number, makes CARP mostly undeployable for ISP datacenter environments,
> and hence there is an obvious need for a working VRRP implementation, it
> doesn't help that CARP is now available in FreeBSD, because it is not a
> viable alternative in a lot of scenarios.
>
As it seems you know how the IETF and IANA process works to, you could go
and get a protocol number for carp and solve the problem. But as usual it
is far easier to come up with some semi-true statements, fallacies and
straw mens than to acctually start fixing stuff.
Now I should shut up and start hacking.
--
:wq Claudio
More information about the freebsd-net
mailing list