ietf-rats / ietf-corim-cddl

This repository is abandoned. The adopted I-D can be found at:
https://github.com/ietf-rats-wg/draft-ietf-rats-corim/
2 stars 0 forks source link

linked-tags rel type 'comid-supplements' semantics #2

Closed nedmsmith closed 3 years ago

nedmsmith commented 3 years ago

The 'comid-supplements' relaltionship type, when used for CoMID to CoSWID, may need to define matching criteria. In comid to comid the comid tag MUST? have a claim with an element-name that matches the element-name of the linked tag. In comid the module-name includes element-name making this matching easy. In coswid there isn't a module-name. Instead, the concise-swid-tag has a set of attributes that need to be mapped to element-name to form the matching criteria.

nedmsmith commented 3 years ago

The other comid relationship types seems to not make sense for comid to coswid linking.

nedmsmith commented 3 years ago

If coswid tag-id can describe a single software item, then linking by tag-id only is sufficient. But if there could be multiple versions / packages referenced by the same tag-id. Then possibly more specific 'matching' semantics may be needed. We think mapping attributes of element-name to attributes in swid could solve this problem.

thomas-fossati commented 3 years ago

This is the certification use case that Yogesh has brought up: certification (e.g., Common Criteria, PSACertified, etc.) metadata associated with a HW module (CoMID) as well as SW module (CoSWID).

thomas-fossati commented 3 years ago

There seems to be agreement at this point in time that the certification use case can be solved by other means, for example introducing a "concise certification tag (concert)" - Henk's mint - to the $concise-tag-type-choice.

I don't remember whether we also concluded that the supplement semantics from CoMID to CoSWID is not supported by enough concrete use cases to be worth adding at this point. @nedmsmith @yogeshbdeshpande ?

yogeshbdeshpande commented 3 years ago

@thomas-fossati : We agreed to the extent as below: 1) As the CoMID to CoSWID supplements does not have a clean solution as far as Certification so as we better remove the use case from the specification. As far as I recollect no other use case requires the cross linking via supplements. 2) We can model the use case by minting two tag, one CoMID that links via Supplements to root CoMID which it certifies. Other is a new CoSWID tag which supplements the root CoSWID tag. The created CoMID tag and new CoSWID tag may contain identical claim set. 3) Follow a common con-cert-tag as suggested above. This may be adapted for specific profiles of standard, as this is not part of the common specifications.

thomas-fossati commented 3 years ago
  1. As far as I recollect no other use case requires the cross linking via supplements.

So we could drop it altogether?

yogeshbdeshpande commented 3 years ago

Since the relationship is common between CoMID to CoM/SW/ID, it only needs an edit in the Endorsement Architecture document, unless I am missing something!

nedmsmith commented 3 years ago

The CoMID to CoSWID supplements case can be removed from the TCG spec.

yogeshbdeshpande commented 3 years ago

Thanks and Agree!, Once the edit is done, we can close this issue!

nedmsmith commented 3 years ago

Removed as of revision 15.

nedmsmith commented 3 years ago

Closed issue based on comments.