[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Comments on draft-ietf-ipsec-ike-01.txt (long)
I have to agree with Michael. Since it costs no more to do 448bit than 128bit
Blowfish, anyone who configures less than 448bits must have a reason. In the
example below, there is no obvious cryptographic or computational reason for
"you" to configure 128bit Blowfish. A reason that an organization might set the
policy in such a manner could be that they can crack 128bit Blowfish and want to
keep an eye on you. Or maybe there is some stupid reason like they use IDEA and
want key escrow and have a broken system that can only store 128 bit keys. It
doesn't really matter. The fact is, some authoritative administrator set the
policy and an implementation should obey the policy. Leaving it a SHOULD
allows the implementater to decide how it will behave. This leaves the door
open for simple implementations that provide straight forward policy handling.
It also allows for complex implementations that allow for "key length or
greater" policies, along with the auditing systems to show administrators what
key lengths users are accepting.
Changing the wording to a MUST forces people to accept key lengths they may not
want, for whatever reason. It also forces implementations wishing to provide
their customers with authoritative policy control to "break" with the standard.
I firmly believe the wording must remain a SHOULD. I also would not like to
see another globally required policy knob that makes people choose between "I
meant what I said" and "You figure it out." Implementations are already free to
contain such things under the current wording.
-Rob
> -----Original Message-----
> From: owner-ipsec@lists.tislabs.com
> [mailto:owner-ipsec@lists.tislabs.com]On Behalf Of Michael C. Richardson
> Sent: Monday, June 14, 1999 1:55 PM
> To: ipsec@lists.tislabs.com
> Subject: Re: Comments on draft-ietf-ipsec-ike-01.txt (long)
>
>
>
> >>>>> "Sandy" == Sandy Harris <sandy.harris@sympatico.ca> writes:
> Sandy> My query would have been whether they should be prohibited from
> Sandy> REJECTING a key length greater than they've configured.
>
> Sandy> You configure for, say 128-bit Blowfish. I offer 448. The
> Sandy> algorithm costs no more to run with the longer key. Clearly you
> Sandy> SHOULD accept. I'd like to see the standard say you MUST accept.
>
> I think the text should say SHOULD.
> Despite Blowfish being able to do flexible key lengths, not all hardware
> may be configured to do that.
>
> :!mcr!: | Network and security consulting/contract programming
> Michael Richardson | ...working from my front lawn with a long cord...
> Personal:
> http://www.sandelman.ottawa.on.ca/People/Michael_Richardson/Bio.html
> Corporate: http://www.sandelman.ottawa.on.ca/SSW/
> ON HUMILITY: To err is human, to moo bovine.
>
>
>
>
>
Follow-Ups:
References: