[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [sig-policy] prop-045: Proposal to modify "end site"definition and allow end sites to receive IPv6 allocations
Hi Jordi,
On Jan 24, 2007, at 6:34 PM, JORDI PALET MARTINEZ wrote:
[...]
I agree that making address space available to networks that need it
is good. My concern isn't the goal but the phrasing of the policy. I
think the proposal is phrased in a way that would allow almost any
organisation where one department 'bills' another for a service to
qualify for a /32 allocation. You implied that the membership fees
are the main deterrent. As the APNIC Fees WG is reviewing the fee
structure I think it might not be a good idea to rely on it when
designing policy.
I think it should be on the other way around. I guess the fees
can't be
changed each time a new policy is passed, but the fees should make
sure that
"in general", fees are charged in a fair way in order to cover APNIC
costs+budget provision for continued operation for a certain number of
years, etc.
So policies should be "fee agnostic", but somehow fees should be
coherent.
My point isn't that the fee schedule will change radically but that
if the policy uses a particular aspect of the fee schedule it may
find that things work quite differently if the fee schedule does
change. For that reason I would suggest aiming for a policy that does
not rely on fees to stimulate or limit demand.
What I meant is that even if the cost is low, not all the companies
will go
for this addressing space, but just in case, fee structure should
consider
this. I guess the questions are should an ISP which is making
business pay
more for addresses than an enterprise which is using those for its own
infrastructure ? Or the other way around ? Should this space
considered a
luxury and thus need to pay more ? Or is a need and should pay
less ? Or
should the cost reflect somehow the extra "slot" in the routing
tables which
in this case is for a single company (while the one for an ISP is
for many)
and consequently be more expensive ? Should the cost be used to
discourage
using extra routing slots/addressing blocks when it is a need ?.
And of course, we can keep moving on that direction, but it will be a
totally different topic and probably scope of the fees WG ...
I think discussion of fees should take place in the Fees WG. I don't
want to argue for any particular model for the fees, though.
I think your goal is to ensure that multi-site enterprise networks
can receive portable address space from APNIC. If that is the case,
maybe it would be useful to ensure that 'site' is less vaguely
defined and change the policy to allow multi-site networks under
common ownership and operation to qualify for an allocation?
It depends on what we consider is a multi-site network. A
university may be
split in many buildings in a single campus, or many buildings in
several
campuses. I think the policy should cover both cases.
I think you are conflating 'enterprise' and 'site'. I agree that a
single enterprise might comprise multiple sites. It might be
appropriate for (some?) multi-site enterprises to receive a /32
allocation. My concern is that the current definition of 'site' is
vague and this makes it difficult to amend the policy in the way you
want. If 'site' is clearly defined it will make it much easier to
write policy text that is really clear on who qualifies for a /32
allocation.
Regards,
--
Leo Vegoda
IANA Numbers Liaison
