[arin-ppml] Draft Policy 2010-2: /24 End User Minimum Assignment Unit - Correct Title
jim at tgasolutions.com
Fri Jan 29 19:46:59 EST 2010
A few points here:
1. If an end user gets a /24 IP address space from an ISP via the /24 MH rule--
Why such a big difference in PA from ARIN?
Yes, I understand-- portable, and not enforced by ARIN for usage.
BUT does the ISP check usage?
ARIN will enforce tighter rules for themselves than an ISP will?
Is IP space usage, regardless of issuing organization, still usage?
If I get a /24 from an ISP, and only use /26 of it, is that any better than the same from ARIN?
Also what about route table growth? See below..
2. If I am going to get a /24 to multi-home for an end user, they will use at least 64.
BUT do we have a formal definition of usage? See below..
Please correct me if I am wrong here.. BUT isn't there a requirement now for a C Level or VP
To put their John Hancock on the request? For Fraud to be a reason to vote this down is wrong. IMHO
Attempts to make rules to account for fraud can never include all the variables to be worth the time.
I have never seen a clear definition of usage. Is it publicly accessible devices? Or NAT pools?
Or association with an IP? Or just someone provide host names to X system?
5. De-aggregation-- If I am an ISP today and I have a /16. And I have to de-aggregate a block of say /19 to give a /24 for an end user..
How much growth do I add to the global route table ? how much if a block of say /16 is used by ARIN to give /24's to end users?
What is the net effect for the global route table? I suspect that "explosive" growth from an ISP dropping the aggregation would be far worse
Than the issuing of IP PA space by ARIN even if an end user is only going to 100% use a /26......
Isn't this route table growth the biggest concern? Do we expect a land rush for IP space if this is passed?
From an ISP perspective, what would you rather do? Announce 10 IP blocks in varying lengths because of 1 end user? Or have that 1 end user
Get PA space and you announce 2 blocks (1 aggregate and 1 /24)
6. End-user requirements for the space..
On the customers that I have gotten /22s for, we were required to provide reverse DNS for the space--
I would assume even a /24 require the same.
With this requirement, won't this help to flush out those that really don't have the ability to handle a /24?
A couple of real world problems I see today:
1. Small ISP-- has a /22 today-- has used approx 65%.
Their space is segmented in such a way where there is not space to provide a /26 to either end user.
They have 2 downstream customers that want a /24 each--they told me they are going to Multi-Home.
But as a small organization, renumbering twice (1 to go to this ISP, and again for a /24 would be burdensome)
The current ISP, is leaving the area...
They are local cable / wireless service providers... ~100 subscribers give or take each.
The ISP is a non-profit that provides service to organizations in the area..
They asked ARIN for more IP space to help the end users, got shot down.
The end users are too small to get the space from ARIN under today's rules.
So what happens?
The end users get several /27s from the small ISP and can't multihome... and then get forced to renumber later.
What would I hope? The ISP could get that /24 or the end user can sign 2 upstreams and get space from ARIN..--
See the sponsorship thought below...
2. end user-- Wants to Multi-home
Both ISP's charge for the /24 space monthly-- large enough that is becomes a challenge to a justify.
Extensive maintenance for both ISPs to de-aggregate the block for multihoming.
(ever watch the CIDR report and check the stats for prefix count increase?--
I watched one local ISP when I multihomed a customer on their IP space.. they jumped up to #6)
(today this end user has a /28 from one provider and a /29 from another, PAT is rampant, route-map routing nightmare)
3. Local small co-op
Has a 2 /25 from a single upstream-- each /25 is from different /24 (bottom of one and top of another)
They have used 60-70%... and want to multi-home...As a power/gas/water co-op they still must get space from an upstream.
Re-numbering will be a lengthy process.... how can we handle this?
Last and my parting thought for the evening--
Sponsorship---And this is just a thought----
It seems to me that ISP's will give the /24 and be put through the ringer to allow for the de-aggregation of their IP space.
So let's assume that ISP A, and ISP B serve customer Z. when ISP A gives a /24 to the customer, there is the ripple effect.
ISP A must do a maintenance do de-ag their blocks. And coordinate this with UPSTREAM 1a thru 1c. who in conjunction may have to do the same to upstream 2a to 2X. and so on. If ISP A and/or ISP B sponsors customer Z and in effect adds comments to the customer's request for space, this would help EVERYONE in the internet community. In this case, the ARIN assigned PA /24 to the end user and it is added BGP filters for the upstreams and outbound from ISP A and B. Seems simpler.... But I know I have over simplified this......
Kind of like a pay it forward for ISPs....
To continue this thought--- Who would know the end user better than the ISP account manager that signed the customer up? So if the sponsorship is made part of the ARIN PA space request, then ISP gets a say in the assignment.-- IE-- simple checkbox-- DOES ISP A support the assignment of a PA /24 to CUSTOMER Z ? Can ISP A provide any documentation to assist or support the need for the /24 from ARIN?
From: wherrin at gmail.com [mailto:wherrin at gmail.com] On Behalf Of William Herrin
Sent: Friday, January 29, 2010 2:57 PM
To: Owen DeLong
Cc: Jim McBurnett; arin ppml
Subject: Re: [arin-ppml] Draft Policy 2010-2: /24 End User Minimum Assignment Unit - Correct Title
On Fri, Jan 29, 2010 at 2:13 PM, Owen DeLong <owen at delong.com> wrote:
> As I understand the NRPM, this proposal would enable customers who can justify
> a need for 128 or more IP addresses and who are multihomed to get a /24. It would
> not enable customers with, say, 8 addresses to get a /24 from ARIN, even though
> they can get a /24 from one of their upstream providers.
The governing language is in NRPM 4.3.3 which reads:
* A 25% immediate utilization rate, and
* A 50% utilization rate within one year.
So that's 64 addresses now and an additional 64 within one year.
Except of course, nobody will check in a year to see if you used an
additional 64 addresses.
IMHO, 4.3.3 should probably be tightened up a bit as we approach free
pool depletion, perhaps to something like "50% now and 75% in a year"
or maybe just "50% now" since we probably aren't going to check in
with end-users in a year anyway. Allowing assignments at the /24 level
will help us do that since it would no longer have the effect of
putting the minimum assignment size out of reach for small multihomed
orgs in the "couple hundred people" category.
> Thus, this policy only opens up ARIN PI /24s to a subset of those that can currently
> get PA /24s, and, not everyone. While I would support a policy that had parity with
> the PA /24 policy for direct assignments from ARIN, I think such a policy would be
> far less likely to achieve community consensus at this time.
Concur. Also, I'd like to see whether any problems (e.g. fraud) crop
up at this proposal's level so that they can be corrected before
attempting for full parity with NRPM 126.96.36.199.
I don't anticipate any problems. ARIN's IPv4 process remains arcane
enough to do a solid job keeping those without a genuine need at bay.
William D. Herrin ................ herrin at dirtside.com bill at herrin.us
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004
More information about the ARIN-PPML