Re: [sidr] Key learning procedures in BGPsec?

[Richard Barnes <[e-mail]>]

On Jan 30, 2012, at 11:57 PM, Brian Dickson wrote:

> On Mon, Jan 30, 2012 at 4:52 PM, Stephen Kent <[e-mail]> wrote:
>> At 2:57 PM -0500 1/30/12, Brian Dickson wrote:
>>> 
>>> There are other kinds of encryption as well, which involve shared
>>> keys, or in case of DH, random session keys with neither party
>>> having/needing the other's key material.
>> 
>> 
>> not true. DH key agreement requires that each party receive the public
>> key of the other, in order to compute a shared secret.
> 
> In DH, each side generates a random (per-session) private key, and
> computes _a_ public key (which is exchanged), off of that private key
> .
> 
> Here's the thing - the "public key" in DH is not pre-published, nor is
> it re-used. It is a one-use, fire-and-forget object, as are the
> private key and shared key.
> 
> This is different from a published (public) key material a la PKI. PKI
> keys are longer-lived and multiple-use.
> 
> In DH, neither party has the other's key _before_ they start their DH
> exchange. That was my point.


<hat type="pedantic"/>

I do not think that word (DH) means what you think it means
Diffie-Hellman is an algorithm on groups that have hard discrete log problems.  It has nothing to do with key management.
<http://en.wikipedia.org/wiki/Diffie%E2%80%93Hellman_key_exchange>

I think what you mean to say is DH with *ephemeral* keys.  That is not a necessary property.  You can have even DH keys in certificates.
<http://tools.ietf.org/html/rfc2631>