Re: [babel] hmac info model elements

["STARK, BARBARA H" <[e-mail]>]

Coming back to modeling of HMAC keys (and PSKs) after the weekend...

Juliusz said: "There's a third model -- locally generated from a passphrase on the local host (e.g. using a PSK)." 
<bhs> If I understand this correctly, this is actually what I thought we would be doing (though I would use the word "derived" instead of "generated", because "derive" to me implies the same result will always be had at the end of the derivation, while "generated" may have a random aspect). Example: WPA (for Wi-Fi, using SHA-1) uses an alphanumeric passphrase / pre-shared key (PSK) (and SSID) that are provided to Wi-Fi endpoints, somehow (various possible methods of providing). The PSK (and SSID) are first converted to Unicode representation of the alphanumeric characters, before being used to derive the HMAC-SHA1 key using the RFC 2898 PBKDF2 pseudorandom function. This is the model I've been envisioning would be used for babel (except we have SHA-256 instead of SHA-1). 

Toke said: "Bird uses an unencoded string..." 
<bhs> I'm not sure what "unencoded" means here? But this sounds like maybe the entered string is a "passphrase" / PSK, like what's used for Wi-Fi, which is then used to derive the HMAC key using Unicode encoding?

Dave wants hex or base64.
<bhs> By this, I'm thinking Dave is actually talking about the HMAC key itself (exactly 256 bits)?

Mahesh said: "RFC8177, YANG Key Chain, models keys as either string or hexadecimal".
<bhs> Now that I've been reading up on HMAC keys -- I think if we're modeling the actual key, it should be hex. If a PSK, then string.

If I'm understanding all of this, I think (?) I'm agreeing with Juliusz and that the info model should support supplying a string PSK, and it should be called something like "hmac-psk" so it's not confused with the HMAC key. We could also allow for supplying the actual HMAC key, which I think should be hex. I'm fine with or without supporting the actual HMAC key in the info model. I think PSK is important.
Barbara

> -----Original Message-----
> From: Juliusz Chroboczek <[e-mail]>
> Sent: Friday, January 04, 2019 5:03 PM
> To: STARK, BARBARA H <[e-mail]>
> Cc: 'Mahesh Jethanandani' <[e-mail]>; Babel at IETF
> <[e-mail]>
> Subject: Re: [babel] hmac info model elements
> 
> >> In my experience, these sorts of shared keys are modeled as strings.
> 
> >> I would think the keys are modeled as binary.
> 
> > That depends on how we generally expect the keys to be generated. If
> > we think they'll be randomly generated by a centralized machine, then
> > binary makes sense. If we think they'll be created by a human being
> > who will input them through a keyboard, then string makes sense. I was
> > thinking the latter would be more common. I'm curious what others think?
> 
> There's a third model -- locally generated from a passphrase on the local host
> (e.g. using a PSK).
> 
> -- Juliusz