RE: [Cfrg] Re: [saag] Algorithm upgrades

["Hallam-Baker, Phillip" <pbaker at verisign.com>]

> On Nov 1, 2004, at 8:09 PM, Hallam-Baker, Phillip wrote:
> 
> >> 1. _New_ protocols that are being design should include the ability
> >> to negotiate algorithms. This is just good hygiene. If this is not 
> >> formalized, it should be.
> >
> > I strongly disagree. History has showed that negotiation mechanisms
> > can lead
> >  to worse security problems than they are meant to address. 
> The IETF 
> > has
> >  certainly showed an exceptional ability to overcomplicate them.
> 
> Well. We disagree then.
> 
> The only history of negotiation failures that I have seen is the GSM 
> vulnerability where they put ECC under weak encryption and then used 
> that same key for high strength, cracked the key, spoofed a base 
> station, mounted a MIM attack and got in. Yes this is a protocol 
> vulnerability with 2 (Doh!) stupid crypto mistakes.

The IETF can start criticizing others for crypto errors after it delivers a
viable spec for ubiquitous IPSEC and DNSSEC. 

Failure to deliver a viable protocol after a decade renders the quality
somewhat meaningless.

> Keeping a protocol centered on the problem without the siren call of  
> "feature creep" is not easy. It is this creep of "just one 
> more feature 
> cuz its a simple tweak" that, when compounded created a 
> incomprehensible stew of incongruous kludges. 

Despite knowing rather more than the average user and several hours of
effort I am unable to get my WiFi card to talk to my main Wifi router with
crypto turned on. 

Apart from SSL I can't think of any IETF protocol that manages a better user
experience.

> Half of the IETF protocols are more complicated than the average IETF 
> protocol.

With the exception of SSL/TLS and WEP virtually all crypto protocols have
failled. WEP also had the minor inconvenience of not working but at least it
is used and the next version will be fixed.

> Well, maybe we strongly agree. What we need is an ability to add new 
> algorithms, remove bad algorithms with causing undue 
> complexity to the 
> user.

What we need is a plan to get people to use the stuff. Real people, not
geeks.

> 
> 
> > What we need is a policy layer for the whole application protocol 
> > layer  stack. It should not be part of the individual protocols, it 
> > should be part of the DNS.
> 
> This seems like a policy punt. what are you suggesting and why?

Nope, recognition of a shortcomming of the Internet architecture. A
principled policy architecture, not the usual punt.

_______________________________________________
Cfrg mailing list
Cfrg at ietf.org
https://www1.ietf.org/mailman/listinfo/cfrg