RJ Atkinson wrote:
- I really don't think that PI allocations are a requirement.
- PI allocations are one possible mechanism to meet a
different requirement.
The underlying requirement is:
The 'pain and cost of renumbering' when multiplied
by the likely frequency of renumbering needs to be
tolerable.
PI space is one mechanism to get that pain and cost down to levels
acceptable to many sites, but it might not be the only way to get pain
and cost down to those levels. We ought not be closing the door on
alternative approaches. Instead, we ought to be encouraging folks to
study and investigate ideas that might lead to other alternative
approaches.
I agree that PI is not an actual requirement, but so far it is the only model which has been shown to meet the true requirements. I'm open to other ideas, if they present themselves in a timely fashion.
Indeed, I can imagine that some of those alternatives might be *lower* cost and *lower* pain than PI space -- particularly given the annual "address leasing" costs associated with getting PI space from a registry.[1]
I don't see how any renumbering system, unless it was 100% automated, could beat PI space on cost. In the ARIN region, the annual "leasing" costs are currently $100 for all direct IPv4, IPv6, and ASN assignments to end user orgs -- one flat fee, regardless of the number or type of assignments. That means a renumbering solution needs to eliminate the need for _all_ of those things _and_ require no more than 1-2 man-hours of labor per year for it to be an economic win.
I'm not saying it's impossible, but it is infeasible with the technology currently available or on the horizon, and IMHO by the time the IETF could develop something and vendors got it to market, PI in IPv6 will be as deeply entrenched as PI in IPv4 already is and few will have reason to switch. Then again, I'm a cynic.
S(I put "leasing" in quotes because numbers cannot be property and thus cannot be leased out; the fees are legally payment for the services of being listed in the RIR's WHOIS and rDNS.)
-- to unsubscribe send a message to rrg-request@psg.com with the word 'unsubscribe' in a single line as the message text body. archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg