Re: [Emu] server unauthenticated provisioning mode

[Glen Zorn <[e-mail]>]

On 8/26/2011 10:56 PM, Dan Harkins wrote:

...

>>>>>   I completely missed that. Sorry. But if the IETF standardized a
>>>>> wholly inappropriate protocol like MSCHAPv2 (it doesn't even generate
>>>>> a
>>>>> shared key)
>>>>
>>>> Please check your sources & refrain from spouting nonsense; if EAP-pwd
>>>> is really so wonderful you shouldn't need to disparage other work, it
>>>> should stand on its own merit.
>>>
>>>   I stand corrected. That must be why draft-zorn-emu-team proposed using
>>> MSCHAPv2. Oh wait, it didn't. It proposed using EAP-pwd.
>>
>> & this has a relation to your misrepresentation  how?
> 
>   It doesn't really, it was a reference to your implication that EAP-pwd
> cannot stand on its own merit, and requires the disparagement of other
> work to prop it up. Which I think we can both agree is not the case.

Actually, Dan, I implied (or intended to imply) no such thing.  My point
was that making demonstrably false claims about other protocols doesn't
strengthen your case but only weakens it, making it appear that EAP-pwd
cannot stand on its own merits.  MS-CHAPv2 was not specifically _not_
standardized by the IETF, the protocol was documented in a purely
informational document ("This memo provides information for the Internet
community. It does not specify an Internet standard of any kind.").
While the ubiquity of Windows undoubtedly had an influence on its
adoption, the main reason IMHO was that it solved a real problem which
was widely disparaged within the IETF as either nonexistent or
inappropriate (i.e., "if you would just change everything you do the way
we told you to, that problem wouldn't exist").

> 
>   Anyway, yes I misrepresented MSCHAPv2. It does produce a key. I was
> wrong.
> 
>   But I still think the statement that MSCHAPv2 is inappropriate for
> doing server unauthenticated provisioning mode is valid even with a
> secret key output by MSCHAPv2. I hope we can both agree on that too.

I am inclined to think that resistance to attack is a valuable thing in
this case; EAP-pwd is demonstrably less susceptible to a variety of
attacks than MS-CHAPv2 (or for that matter, EAP-GPSK).  Furthermore, I
think that constraining the choice of a MTI type on the basis of a
single academic project (instead of security) is basically unconscionable.