[167562] in North American Network Operators' Group
RE: IPv6 /48 advertisements
daemon@ATHENA.MIT.EDU (Eric C. Miller)
Thu Dec 19 00:52:04 2013
From: "Eric C. Miller" <eric@ericheather.com>
To: Owen DeLong <owen@delong.com>
Date: Thu, 19 Dec 2013 05:51:38 +0000
In-Reply-To: <90016801-9EA1-4D33-93C4-A49EF782EC49@delong.com>
Cc: "nanog@nanog.org" <nanog@nanog.org>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
Owen, thanks for this explanation. +1!
Eric Miller, CCNP
Network Engineering Consultant
(407) 257-5115
-----Original Message-----
From: Owen DeLong [mailto:owen@delong.com]=20
Sent: Wednesday, December 18, 2013 2:07 PM
To: Cliff Bowles
Cc: nanog@nanog.org
Subject: Re: IPv6 /48 advertisements
On Dec 18, 2013, at 08:11 , Cliff Bowles <cliff.bowles@apollogrp.edu> wrote=
:
> I accidentally sent this to nanog-request yesterday. I could use some fee=
dback from anyone that can help, please.
>=20
> Question: will carriers accept IPv6 advertisements smaller than /48?
Generally, no. Since a /48 should represent nothing larger than a single si=
te, it's not very reasonable to want to route something longer in general.
> Our org was approved a /36 based on number of locations. The bulk of thos=
e IPs will be in the data centers. As we were chopping up the address space=
, it was determined that the remote campus locations would be fine with a /=
60 per site. (16 networks of /64). There are usually less than 50 people at=
the majority of these locations and only about 10 different functional VLA=
Ns (Voice, Data, Local Services, Wireless, Guest Wireless, etc...).
That's still poor planning, IMHO. You can easily get more than enough /48s =
to give one to each location. There's absolutely no advantage in the IPv6 w=
orld to being stingy with address space and no benefit to not putting at le=
ast a /48 at every location.
You've got 10 VLANs, so you're wasting at most 65,526 networks. Compare tha=
t to the fact that using a /64 for a VLAN with less than 2,000,000 hosts on=
it will wast at least 18,446,744,073,707,551,616 addresses and you begin t=
o realize that sparse addressing in IPv6 and large amounts of excess addres=
s capacity are intentional.
> Now, there has been talk about putting an internet link in every campus r=
ather than back hauling it all to the data centers via MPLS. However, if we=
do this, then would we need a /48 per campus? That is massively wasteful, =
at 65,536 networks per location. Is the /48 requirement set in stone? Will=
any carriers consider longer prefixes?
Massively wasteful is a fact of life in IPv6. Consider it this way... There=
are two ways to waste address space. One way is, as you describe above, de=
ploying it to locations that are unlikely to fully utilize it.
Another way is to leave it sitting in a free pool until long after the prot=
ocol is no longer useful.
With IPv6, we're not so much choosing between wasting address space or not.=
We're choosing how much address space gets wasted using method 1 vs. how m=
uch gets wasted using method 2. Ideally, we arrive at the protocol end of l=
ife with some space remaining in both categories of waste.
> I know some people are always saying that the old mentality of conserving=
space needs to go away, but I was bitten by that IPv4 issue back in the da=
y and have done a few VLSM network overhauls. I'd rather not massively allo=
cate unless it's a requirement.
It's a requirement and not massively allocating will bite you harder in IPv=
6 than space did in IPv4.
IPv4 was designed for a different kind of network. It was designed to suppo=
rt some labs and some institutional environments. It was never intended to =
be the global public internet. IPv6 has been designed with the idea of addr=
essing absolutely everything from the ground up. The design allows for plen=
ty of /48s to number every building that could possibly fit on every planet=
in the solar system and several other solar systems.
Frankly, a /48 per campus is underallocating for any campus that has more t=
han one building.
Owen
