[arin-ppml] Rationale for /22

David Williamson dlw+arin at tellme.com
Tue Jul 28 13:10:26 EDT 2009


Why not?  I've thought of writing a proposal that does exactly that.
The routability of the block isn't nearly as interesting as good
stewardship of the available space.  If all you need is a /26, why get
a /24 (or, under current rules, a /22)?  Sure, it may not be trivially
routable today, but that's a limitation of the routing system, not an
excuse to burn space.

At some point, the routing system will need to have a way to identify
the value of a route.  Suppore, hypothetically, windowsupdate.com lived
behind a VIP that is advertised as a /28.  Would you route it?  Of
course!  If you didn't, a competitor would, and every one of your
customers with a Windows box (i.e., most of them) would leave.  Would
you route the /28 for joesautobody.com?  Probably not.  How do you identify
the routes you find valuable and are willing to burn a routing slot for?
Well, that's the trick...no easy answers to that one.

The current /24 "limit" is an artifact of the current situation.  The
limit has been different in the past.  As recently as 5 years ago, it
was /22 for some major providers.  It'll slide to /25 at some point.
Let's not enshrine that in policy, except perhaps as a stepping stone
to the no-limit world.

-David


On Tue, Jul 28, 2009 at 12:34:55PM -0400, Steven E. Petty wrote:
> 
>   Obviously the primary reason for a limit is the accepted minimum routing entry size of a /24.  Of course, you knew that.  So why ask the question?
> 
> 
> -----Original Message-----
> From: arin-ppml-bounces at arin.net [mailto:arin-ppml-bounces at arin.net] On Behalf Of Kevin Kargel
> Sent: Tuesday, July 28, 2009 12:30 PM
> To: William Herrin; Joel Jaeggli
> Cc: ARIN PPML
> Subject: Re: [arin-ppml] Rationale for /22
> 
> Based on what everyone is saying about the /24 issue - and for the purpose of argument accepting that /24 will cause no problems - then why not take it a step further and remove any maximum length netmask restriction for a multi-homed entity with a single allocation.
> 
> One entity with one allocation will generate one table entry regardless of what size it is, so why limit them to a /24.  I am sure there are entities out there who could operate just fine out of a /25 or even a /32, and so long as they are not creating gratuitous route table entries then we could be more efficient allowing them to only consume the space they need.
> 
> 
> 
> 
> > -----Original Message-----
> > From: arin-ppml-bounces at arin.net [mailto:arin-ppml-bounces at arin.net]
> > On Behalf Of William Herrin
> > Sent: Tuesday, July 28, 2009 10:30 AM
> > To: Joel Jaeggli
> > Cc: ARIN PPML
> > Subject: Re: [arin-ppml] Rationale for /22
> >
> > On Tue, Jul 28, 2009 at 10:32 AM, Joel Jaeggli<joelja at bogus.com> wrote:
> > > William Herrin wrote:
> > >> Given the shortage of IPv4 addresses, why structure the policies so
> > >> that we give anyone more than they actually want?
> > >
> > > The minimum number of addresses that can be used may not in fact
> > > reflect the minimum that should be used.
> > >
> > >        For the purposes of minimizing fragmention.
> > >
> > >        Supporting basic network operation (it's nice when traceroute
> > >        and pmtud work) because your intermediate routers are privately
> > >        numbered.
> > >
> > >        Limiting the consequences of imagination failure, which may
> > >        sound flippant but renumbering, requesting an additional block,
> > >        or and point one and two are good reasons to make a potential
> > >        multi-homer justify the assignment of a block of the appropriate
> > >        size for that activity.
> >
> > Hi Joel,
> >
> > I could almost see that argument on the ISP side but it doesn't make
> > sense to me on the end-user side, particularly when they may be
> > trivially multihomed. I surely wouldn't want to presume that I know
> > every registrant's address count needs better than he does though. f
> > you don't mind, let's just focus on the downside risk.
> >
> > So, the registrant asks for a /24 and a year later his replacement who
> > is a better network engineer figures out he really needs a /22 after
> > all. What's the impact? Other than insisting on giving him a /22 up
> > front, is there another way to mitigate that impact?
> >
> > Regards,
> > Bill Herrin
> >
> >
> > --
> > William D. Herrin ................ herrin at dirtside.com  bill at herrin.us
> > 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
> > Falls Church, VA 22042-3004
> > _______________________________________________
> > PPML
> > You are receiving this message because you are subscribed to the ARIN
> > Public Policy Mailing List (ARIN-PPML at arin.net).
> > Unsubscribe or manage your mailing list subscription at:
> > http://lists.arin.net/mailman/listinfo/arin-ppml
> > Please contact info at arin.net if you experience any issues.
> _______________________________________________
> PPML
> You are receiving this message because you are subscribed to
> the ARIN Public Policy Mailing List (ARIN-PPML at arin.net).
> Unsubscribe or manage your mailing list subscription at:
> http://lists.arin.net/mailman/listinfo/arin-ppml
> Please contact info at arin.net if you experience any issues.



More information about the ARIN-PPML mailing list