Bad routes update

[Robert Rockell]

->Right now, there don't seem to be any real restrictions on the
->practice in v4 land other than taste, and yet, we aren't overwhelmed
->yet. I suspect that the desire will not expand faster than Moore's
->Law. If it does, market (financial) rather than technical solutions
->would seem to me to be superior.

I disagree. With registries demanding the providers delegate down below /24
and /25, we can begin to expand more exponentially than may be first noted.
With singly-homed (at least to one provider) sites, we may not see this
unscale too quickly. however, with multiple-provider "small network"
customers, we can imagine routing tables exploding beyond what we expect.
Remember, small networks may be large networks with firewalls, so this can,
in affect, get popular fast, if there are more address space
conservationalists out there.

As far as letting multi-homing go with v6 as it appears people do now, 
I can see the value in that, but I believe it is important to keep our
engineering/operator hats on. To see multi-homing work correctly (one option
is via RFC2260; outside of whether or not it will work), I think it
important to work early on this, instead of the end-around approach we have
had to suffer through with IPv4 (IPSec, traffic engineering, Qos, routing
table size, etc...) I wonder if it important to keep the tenaments of IPv6
that we have held important since the protocol's inception, and work now to
fix the problems and retain the inherent benefits the protocol can give,
rather than compromise protocol beauty for "what will hopefully scale".
Moore's Law MAY have an upper limit (though maybe not a forseeable one). 

Thanks
Rob Rockell
Sprintlink Internet Service Center
Operations Engineering
703-689-6322
1-800-724-3329, PIN 385-8833
Ines|e gnyne qh vagr bz s|e Ino ngg una {e hgr bpu plxyne?

On 20 Jul 1999, Perry E. Metzger wrote:

->
->Magnus Ahltorp <map@stacken.kth.se> writes:
->> >  o  stability of individual sessions is not important in the event of
->> >     a routing change upstream;
->> 
->> There is an internet-draft on how to use Mobile IP to solve this
->> issue.
->
->And how much operational experience is there?
->
->> Do you really think that there are only 80000 sites that want to be
->> multi-homed?
->
->Right now, there don't seem to be any real restrictions on the
->practice in v4 land other than taste, and yet, we aren't overwhelmed
->yet. I suspect that the desire will not expand faster than Moore's
->Law. If it does, market (financial) rather than technical solutions
->would seem to me to be superior.
->
->Perry
->