> Could you define "legitimately multihomed" please?
... or any other high
This would be a semi-justfiable use of a portable block
less than a /19, would it not?
It would if it could be defined.
Speaking for myself (not ARIN or its members) I would say yes, however,
it's easy to list examples of multi-homed but we would need to either
list every acceptable example or state one definition that doesn't
make ARIN's staff have to subjectively decide which organization is
"legitimately" multi-homed. From my experience (not to mention ARIN's
Advisory Council) this is more difficult than it seems.
ARIN and others including the AC have struggled over what this definition
is. Easily enough if we specify minimum bandwidth or cost associated
with multihoming, but even this is not universially agreed on,
and I'm positive that someone can poke holes in it and claim they
need PI space even when they don't meet the qualifications. The
simple problems have been solved. If it isn't fixed, it must be hard.
I'd say any organization with multiple 100mbps NAP/MAE connections
interconnected at DS3 level obviously merits PI space(but no more
than they can justify). As to how much space someone that is
at one or two exchange points, and a couple of T-1s, people will argue
about. Then what about 2 56k lines? (Some will claim you can't
run two full BGP feeds off less than dual T1s, but I've done the
tests to prove otherwise.).
People also forget that there have been informal talks between
sprint people and registry people about relaxing filters to match
Frankly I don't see policy changing, unless people provide
helpful feedback to the registries and advisory councils. And when providing
feedback its important to be clear and to the point, a lot of these
people are very busy and can't bother with personal attacks and flames.
Its already to late to make the agenda for the Oct 15th AC meeting.