Thus spake "Roger Jorgensen" <[EMAIL PROTECTED]>
On Sat, 7 Jul 2007, Stephen Sprunk wrote:
Thus spake "Roger Jorgensen" <[EMAIL PROTECTED]>
(we are looking at something between 500 and upto 10 000 or
more unique ULA-C/G blocks for or network and use)
That's going to get expensive, no matter how little the central
authority charges. Well, unless the central authority does bulk
discounts, but I see no reason they would since there's no
economy of scale in issuing thousands of prefixes that can't
be aggregated, unlike aggregatable PI/PA space.
Did you notice I didnt mention the price, cheap or not one time
in my mail? We know it will cost so we dont consider that yet.
Make ULA-C/G more expensive than PI and we likely will use
PI, make PI more expensive and it is not an options. You see
the price dont mather! It is the technical options and
possibilities that are important.
I've seen three reasons for ULA-C/G proposed:
1. People can't get PI space
- This shouldn't apply to you.
2. ULA-C/G will be cheaper than PI space.
- This remains to be seen.
3. ULA-C/G won't be publicly routed.
- There's no requirement you route PI space.
and we do NOT care about aggregation much either since it
is our own internal network we are talking about.
You don't do any route summarization within your internal network? Must be
nice not to have to worry about that.
I do not see a requirement to pass ULA-C/G merely because
some RIRs have failed to meet their respective communities'
needs. If you're unhappy with what RIPE offers, either show
up and propose policy changes or (less-preferred) route
around the failure and get space from ARIN.
you totaly missread what I wrote. A PI policy will most likely be
in place in RIPE region to at some point soon, thats why I said
we could choice from PI or ULA-C/G.
So, then, the question is what does ULA-C/G buy you that PI does not?
Did you also notice that I said we have already got our own PA
assignment, a /32, for all our external needs?
Okay, so you already have more addresses than you could possibly need. When
you said PA, I assumed you had a typical /48 block(s) from your upstream(s)
and that wasn't enough space and/or you wanted provider independence that
you didn't have today.
Almost all ULA-C/G usecases can be solved by using either PA
or PI type of addresses from the RIRs but why not give the
enterprises an options to use something similar to RFC1918
addresses?
We already have something similar to RFC 1918: RFC 4193.
ULA-C/G is _not_ similar to RFC 1918; it is PI by another name.
Thats what we are talking about, another set of addresses for
inter-network use, NOT DFZ/global Internet usage.
Addresses are addresses; you can use PI space for private use just as easily
as, if not more easily than, you can use any other type of space.
S
Stephen Sprunk "Those people who think they know everything
CCIE #3723 are a great annoyance to those of us who do."
K5SSS --Isaac Asimov
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
