Re: [v6ops] siit-eam/siit-dc vs IVI

[Tore Anderson <[e-mail]>]

Hi Tariq,

(I changed the subject so that I can easily find this message later.)

* Tariq Saraj <[e-mail]>

> I don't understand the reason for discussion on SIIT when IVI is
> there. The flaws are addressed and mitigated in IVI.

See the Problem Statement section in SIIT-EAM:

http://tools.ietf.org/html/draft-anderson-v6ops-siit-eam-01#section-2

While IVI doesn't use IPv4-translatable IPv6 addresses, it uses its own
address format which share all the disadvantages of IPv4-translatable
IPv6 addresses. As specified in RFC 6219 section 3.1:


     | 0                 |32 |40                   |72             127|
     ------------------------------------------------------------------
     |                   |ff |                     |                  |
     ------------------------------------------------------------------
     |<-     PREFIX        ->|<-  IPv4 address   ->|   <- SUFFIX ->   |

The key problem with inflexible formats such as this one is that it
require that the operator must design and number his IPv6
infrastructure specifically in order to facilitate IVI translation.
This makes IVI simple at the expense of making the IPv6 infrastructure
complex.

SIIT-EAM (and by extension SIIT-DC) allows for mapping between
arbitrary IPv4 and IPv6 addresses; no particular address format is
mandated. So with SIIT-EAM you could have an IPv4 address 192.0.2.1 map
to the IPv6 address 2001:db8::f00. That is, as far as I can tell,
impossible to accomplish with IVI (in IVI, 192.0.2.1 would map to
2001:db8:ffc0:2:100:: as I understand it).

Another key point is that IVI (RFC6219) isn't a proposed standard, so
you can't implement IVI with a standard SIIT implementation. That's
true for SIIT-DC as well, and it is something I attempt to fix with
SIIT-EAM. That said, SIIT-EAM in its current form cannot be used to
implement IVI, because it forbids the use of differing suffix lengths:

   An EAM's IPv4 Prefix and IPv6 Prefix MUST have identical suffix
   lengths.  Any suffix bits MUST be kept intact during translation.

This conflicts with IVI, which allows for different suffix lengths (and
fixing up any discrepancy by padding any missing suffix bits with zero):

   Note that based on the IVI mapping mechanism, an IPv4 /24 is mapped
   to an IPv6 /64, and an IPv4 /32 is mapped to an IPv6 /72.

It should be possible to change SIIT-EAM such that it would be possible
to express IVI mappings with it too. Would you find that valuable?

Tore