[ppml] 4.4.2 Micro-allocations for anycast services

Owen DeLong owen at delong.com
Tue Apr 11 10:00:11 EDT 2006



--On April 11, 2006 9:26:51 AM -0400 Scott Leibrand 
<sleibrand at internap.com> wrote:

> On 04/10/06 at 11:17pm -0700, Bill Woodcock <woody at pch.net> wrote:
>
>>       On Mon, 10 Apr 2006, Scott Leibrand wrote:
>>     > I have an unanswered question regarding anycast, namely: why do
>>     > you need a micro-allocation to do it?
>>
>> You don't.  However, if all you need is one IP address, a microallocation
>> is less wasteful than a larger block.
>
> True, but no more less wasteful than using a /24 out of a larger block (if
> you have one).
>
However, because at least part of the intent of many anycast implementations
is to overcome DDoS, it is at least desirable in many of those cases to
be able to put your anycast stuff in a block that is not directly associated
with your existing infrastructure.

>>     > > At present, the minimum allocation requirements pose an obstacle
>>     > > to smaller to medium DNS providers who want to test and deploy
>>     > > anycast services.
>>     >
>>     > Can you elaborate on this?  I don't see reachability obstacles, as
>>     > stated above.  What other obstacles are there to using your
>>     > existing space to do anycast?
>>
>> Exactly what he said: the minimum allocation requirement.  If all you
>> need is one IP address, justifying a /22 requires the making of, um,
>> unsubstantiatable assertions, to ARIN analysts.  That's an unfortunate
>> state of affairs, and one which this policy will correct.
>
> Ok.  If this policy is intended for allocations to organizations who don't
> have PI space, I don't have any real objections to it.  I would be
> somewhat more comfortable if we explicitly stated the expectation that if
> you have PI space you should use it for this purpose, and only be eligible
> for a micro-allocation if you're not big enough to need a PI /22.  But I
> won't try to hold up the policy until the next meeting based on that
> quibble.
>
Just out of curiosity, does the information above address your quibble?

Owen


-- 
If this message was not signed with gpg key 0FE2AA3D, it's probably
a forgery.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20060411/45990294/attachment.sig>


More information about the ARIN-PPML mailing list