[rrg] Fundamental objections to a host-based scalable routing solution

[William Herrin]

On Mon, Nov 24, 2008 at 3:04 PM, Tony Li <tony.li at tony.li> wrote:
> |The good news is that the RRG doesn't get to decide
> |about this. A host based solution will or won't deploy,
> |and will or won't succeed, on its own merits and regardless
> |of both the routing community and the ISPs. So I don't
> |think we need to discuss it here at all.
>
>
> The bad news is that we have to discuss this.  If a host based solution is
> in fact undeployable, then there's little point in RRG coming to consensus
> on one.

At least one train of thought suggests that a host-based solution is
more easily deployed than a network-based one. The rationale goes
something like this:

1. IPv6 was supposed to be a phased deployment: a. Install the
software ubiquitously. b. Turn it on. c. Displace IPv4 as the primary
carrier. d. Turn IPv4 off.

2. We completed step a. Why suspect we can't do it again?

3. IPv6 deployment died at step b. There's no reason to suspect other
network-based strategies won't suffer the same fate.

4. If a host-based protocol allows us to use existing layer-3
protocols during step c, we can move step b after step c, improving
step b's chance of being doable.

Correct? Maybe, maybe not. I'm far from certain that its not.

Regards,
Bill Herrin


-- 
William D. Herrin ................ herrin at dirtside.com  bill at herrin.us
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004