[ppml] HD Ratio changes

Michael.Dillon at radianz.com Michael.Dillon at radianz.com
Wed Feb 11 06:24:44 EST 2004

>The HD Ratio Policy seems to be getting complicated.  We should keep
>things SIMPLE. 

4 sentences...

>Especially since there will be newcomers to the ARIN
>policies that will need to learn how to adopt our practices in a short
>period of time.

The solution to that problem is better education.
That's why I have posted links to other documents
and why I am preparing some more detailed examples
including charts and tables that will be posted to
the web once this policy proposal has been formally

>Can it be said that the rediscovery and repair of previous allocations
>inefficiencies (or rather their inclusion towards current utilization
>rates) could prove more fruitful in the quest for IPv4 address space? 

Every network is different. But under the current
policy it would be very unwise for an ISP to fix
previous inefficiencies. Instead, they should only
use addresses from their most recent allocation and
only resort to older addresses if they run into shortages
after having exhausted the last allocation and before
getting a new allocation approved. The current policy
discourages ISPs from recovering and reusing address
space. Those older addresses represent flexibility
while the 80% threshold represents rigidity.

>If a new policy is ratified, will it only apply to new allocations?

That's what it says. And it is optional too. If you
can't figure it out, then stick with 80%. If your tool
vendor hasn't implemented the HD Ratio then stick with
80%. If you do complex analyses and comparisons of
the HD Ratio and it works out worse than 80% then stick
with 80%.

> Practices that would include the (true) inclusion of
> previously allocated address space, and the more meticulous weeding out
> of 'ghost' reassignments (SWIP's of customers that are no longer there).

This new policy encourages this to happen because it 
no longer forces larger networks to do the impossible
and because it accomodates the need for longer lead 
times that are common in larger organizations. Today
people have to hide a stash of ghost reassignments to
use as a buffer against the uncertainty of the ARIN
allocation process and the rigidity of the 80% threshhold.

>Rather than rehash an existing policy, I think we should look elsewhere
>for NEW ways to extend the life of our IPv4 address space, both old and
>new. (And enforce those measures).

I'm not interested in extending the life of our IPv4 address 
space. We have twenty years supply left and if the HD ratio
cuts that down to 18 years, then I'm still happy. In about
10 years I expect that IPv6 deployment will be so widespread 
that people will start returning IPv4 space. In other words
I believe the IPv4 address space will never run out.

--Michael Dillon

More information about the ARIN-PPML mailing list