On Fri, 13 Jul 2007, Stephen Sprunk wrote:
Thus spake "Roger Jorgensen" <[EMAIL PROTECTED]>
On Thu, 12 Jul 2007, Stephen Sprunk wrote:
This has been a longstanding problem in the IETF; in fact,
the inability to agree on what "site" means was one of the
reasons SLAs were deprecated. The word "site" is often
abused to mean "administrative domain" rather than "physical location" due
to the ISP-centric nature of the IETF
and RIRs. It's virtually impossible to tell, in any particular
context, which meaning an author meant. We need to stop
using the word entirely...
why does it mather if it is physical site, administrative domain
(eh how can anyone interprent that as a site?!) or a room,
company etc... why not leave that upto those that try to use it?
It matters because there are a lot of protocol and policy documents that
refer to "sites". For instance, one of the current qualifications for
getting an IPv6 LIR allocation require that one have a plan to make
assignments to 200 "sites". Obviously the meaning of that is significantly
changed depending on whether you consider a single company with 200+
locations sharing a single uplink to be one "site" or 200+ "sites".
if you refair to RIPEs policy it has just been changed,
"Consensus has been reached on the proposal described in 2006-02 and
it is accepted by the RIPE community.
This proposal is to change the IPv6 Initial Allocation criteria and
the End Site definition in the "IPv6 Address Allocation and
Assignment Policy".
You can find the full proposal at:
http://www.ripe.net/ripe/policies/proposals/2006-02.html
"
Also, in the case of SLAs, the utility of the addresses is impacted greatly
by whether you consider a "site" to be a single administrative domain, where
there would not be internal collisions, vs. considering each distinct
location to be a "site", which could lead to hundreds or thousands of
internal collisions.
I can not see how SLAs are any of this wg's business, or ietfs businness.
thought SLA was something for the laywers that write the contracts
governing any SLAs to fight out between them?
As far as why "site" has been abused to mean "administrative domain", that
comes from the IETF and RIRs being very ISP-centric, as I said; a single
downstream connection denotes a single "site" regardless of how complex the
internal network behind it is or how many other locations it serves. Or
maybe it doesn't, depending on who's talking; that's the problem.
that downstream connection is ONE site for the upstream provider, will
probably never be anything else. Why should it be anything else really? If
the same "site" bought two connection to an ISP on different location they
would still be one huge site, with several internal parts, two that the
ISP need to know about.
What I'm not sure you realize is that the moment anyone define a site to
be something very specific... there will come alot of exceptions to that
list, not to forget discussion on how to interprent whats written there.
It is way easier to say site and leave it with that... and maybe, just
maybe list a few examples and leave it upto those that have to use the
word site to agree between them.
--
------------------------------
Roger Jorgensen | - ROJO9-RIPE - RJ85P-NORID
[EMAIL PROTECTED] | - IPv6 is The Key!
-------------------------------------------------------
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
