Re: [IPsec] revisiting 3DES and -graveyard

[Robert Moskowitz <[e-mail]>]

Just as an aside thought about 3DES:

perhaps you saw my questions to the CFRG list where I have exactly 64 
bits to encrypt and no place for an IV or such.

One of the serious suggestions WAS 3DES with 3 keys.

For a number of reasons I am not offering that in the initial ID, rather 
AES-CFB16...

But for only 64 bits to encrypt, 3DES is a consideration.

Nah.

(also it may change to exactly 96 bits to encrypt.  They left out the UA 
Altitude and the FAA is not happy with that).

Have a good weekend all!

On 4/15/20 8:49 AM, Tero Kivinen wrote:
> Benjamin Kaduk writes:
>> I see in
>> https://datatracker.ietf.org/meeting/104/materials/minutes-104-ipsecme-00
>> that we didn't want to get rid of 3DES at that time.  Do we have a sense
>> for how quickly that will change, the scope of existing deployments, etc.?
> One of the problems is that we as an IETF give instructions to
> implementors, not to users. If we change 3DES MUST NOT, then all
> implementations out there who do implement 3DES, but where it is
> disabled in configuration by default are no longer conformant to the
> new RFC, as they do still implement 3DES.
>
> Anybody who puts 3DES in any of the new implementations or new
> releases of the current implementations are going against the SHOULD
> NOT in the current RFC.
>
> Meaning they most likely have some old legacy stuff they need to
> support which only supports 3DES and thats why they want to keep 3DES
> in their current releases as they want to be able to talk to those.
>
> I would wait bit more than 2.5 years before saying MUST NOT.
>
>> In particular, would a general-purpose OS's implementation cause problems
>> for its consumers if the next release dropped support?  (Noting that
>> consumers could stay on an old OS release to match the old algorithms, at
>> least for a while.)
> Especially with consumers and general-purpose OS there is no option
> for using old OS release anymore. Most of those will automatically
> update to latest version, and there usually isn't any easy way to
> downgrade back to previous version when issues are found.