Agenda Item: TLS mechanisms for ECDH-type cipher suites

[Dina Kurktchi <dina.kurktchi@oracle.com>]

Hi all,

I'd like to request an agenda item for the next meeting
to discuss adding some new mechanisms in support of other
cipher suites beyond simply 'original' and 'DH' types,
for example:
	CKM_TLS_MASTER_KEY_DERIVE
	CKM_TLS_MASTER_KEY_DERIVE_DH

We're seeing some confusion in consumers of our PKCS#11
implementation about which to use for ECDH type cipher
suites.

It looks like some have chosen the 'original' to derive
master keys because of the "fixed-length" phrase in
PKCS#11 description.  Then others have chosen the 'DH'
type because ECDH is DH-ish, if you will.  In those cases,
we have run afoul of a strip-leading-zeroes problem in
some implementations of DH.

The inclination is to create a new family of TLS mechanisms
specifically for ECDH, to avoid the guessing, such as:
	CKM_TLS_MASTER_KEY_DERIVE_ECDH

I'd like solicit the group's thoughts on this.  Please
feel free to reply and share your opinions.  We can
discuss at next meeting possibly.

Thanks,
D.