Re: [kmip] KMIP Spec v1.2 wd05: Multiple Cryptographic Parameters for a Single Key

[Tim Hudson <tjh@cryptsoft.com>]

On 2/07/2013 3:28 PM, John Leiseboer wrote:

Even if we don't change this in the specification, I don't think that it looks good for KMIP to show insecure examples.

John, KMIP is about demonstrating interoperability and recognising that users of key management products make their own decisions about the context in which they operate and what decisions make sense from a security perspective. We have not to date made decisions to preclude the management of objects where the objects may be used in an insecure manner.

The discussion in the PKCS11 TC which you are quoting is both out of context and also not relevant - it is discussing a new mechanism - and not talking about removing the existing usage at all. No one in that TC is suggesting complete removal of what is actually in active use. Chris is commenting to Mike that if he wants a new mechanism to address that issue he has to also note that the new mechanism has to preclude the older v1.5 format as well as mandate OAEP.

Are you suggesting:
1) we ban the use of ECB?
2) we ban the use of RSA1024?
3) we ban the registration of DES keys?
4) we ban the registration of RC2 keys?
5) we ban the registration of certificates containing RSAwithMD2?
6) we ban the registration of certificates containing RSAwithMD5?
7) we ban the use of PKCSv1.5 padding?

That is just picking a couple of very simple examples of items that can impact security.

These are all choices that the implementers of the products which support KMIP should make and not items that we should (in my view at the very least) be declaring forbidden within the KMIP specification.

As I've been saying for the last two years and others have echoed, the items you are raising are real - but the forum and the context in which you are trying to address them is simply not the right one. If the majority of the TC feel we need to fundamentally change direction then that would be a good discussion to have and would lead to an incredibly large set of changes to KMIP (and corresponding impact on deployed product), however given that to date no one else has expressed anything like the passion you have for tackling these items in KMIP and no one else has express that they think that we should change direction that you accept that the consensus view is simple different to that of your own.

Again - I suggest you define a profile - on top of KMIP - that addresses the concerns you have - and seek support for that profile through the normal process.
The process of going through and defining a profile will help crystallise your thoughts on the topic and also make it clearer the range and scope of changes you would want to see.

Tim.