Minutes of the October 15-16, 1997 Meeting of the Technical Advisory Committee to Develop a Federal Information Processing Standard for the Federal Key Management Infrastructure October 15, 1997 The sixth meeting of the Committee was called to order at 9:10 a.m. by the Chair, Dr. Stephen Kent. In addition to the Chair, members present were: Joe Alexander, Josh Benaloh, David Carman, Santosh Chokhani, Paul Clark, John Edwards, Mark Etzel, Roger French, Russ Housley, Mike Markowitz, and Don Rothwell. Government liaisons in attendance were: Elaine Barker (NIST substitute), Howard Bolden, Dianne Dunshee, Michael Gilmore, Barbara Kirsch, Julie Lever, Jan Manning, and Patricia Sefcik. Ed Roback, Executive Secretary, welcomed everyone and reviewed the agenda for the two day meeting, which was comprised of a series of working group reports and ensuing discussion. (See Reference #1.). As with all TAC meetings, the entire meeting was conducted in open, public session. Dr. Kent's Opening Remarks In his opening remarks, Dr. Kent stressed the need for members to read and comment upon other WG drafts. He also noted that: 1) the examples included in the materials thus far focused more on recovery of transmitted information vice storage; this should be better balanced; 2) the WG #1 and WG #2 sections are not phrased in the manner of "requirements" appropriate for a FIPS; and 3) there is a need to ensure that compliant implementations can be mapped to the overall model. He then provided feedback on each of the WG drafts. The draft of WG #1 is very far along. Dr. Kent suggested that the examples given be tied to the material in WG #5. He again noted the need for the drafts to take on more of a "requirements flavor" and for the inclusion of additional diagrams in the WG #1 section. Even though this standard may not be as detailed as some other FIPS, there needs to be sufficient detail so that non-compliance can be ascertained. With regard to WG #2, Dr. Kent noted that overall the structure of the effort is much better and moving in the direction of requirements but is still at a fairly high level. The security section also needs to move in the direction of "tiering" so it can tie to the assurance requirements developed by WG #8. WG #5's draft is the closest in terms of requirements language that would be appropriate for inclusion in a FIPS. It is very far along, but needs to be closely coordinated with other WG drafts. Consistency in terminology will also require attention. WG #8 is also very far along, and has the right language for assurance specifications. Dr. Kent suggested that WG #2 examine the work of WG #8 and bring in the functionality/security requirements to link with the assurance specifications. The TAC as a whole will have to decide how far to go in terms of bundling the two together. Dr. Kent thanked everyone for getting WG drafts in on time. The agenda then turned to reports from the working group chairmen. WG #1 Report Mr. Roger French, Chair of the Framework Working Group, provided a short summary of the activities of the WG since the last meeting. In response to general comments on the draft: WG #1 will add references to the work of other WGs; ensure the "smoothing" of the text for consistency, level of detail and readability; add a stored data example, and migrate towards "requirements" language. The need for attention to the glossary (currently partially completed in the Announcement section) was also raised. The Committee as a whole then proceeded to review the draft section paragraph by paragraph (not included here). Detailed written comments were also requested by Mr. French. (ACTION - Members) WG #2 Report Dr. Benaloh reported on the progress of WG #2 (Security). (See references #3 and #4.) He reviewed each of the primary security components: general system, end user, key recovery agent, registration agent, and authentic public key source. Each were discussed in detail by the group and critiqued. Various comments were offered. The group is planning to produce a matrix of security / functional requirements at each of the three tiers (corresponding to those of WG #8) for each component of the overall model. In the course of the discussions, references and questions arose regarding the selection of assurance specifications included in the matrix proposed by WG#8. Dr. Chokhani offered to send the committee references to three documents that provide additional details on each. (Action - Dr. Chokhani) WG #5 Report The Interoperability Working Group report was provided by Dr. Paul Clark. (See Reference #6.) Dr. Clark reviewed each of the interoperability requirements. During the discussions, it was suggested that the term "medium" needed to be better defined. Two requirements, #6 and #9, generated considerable discussion. Requirement #6 deals with recoverability of signature keys, which the government wishes to avoid (leading to a likely need for dual key pairs). Requirement #9 states that an algorithm-independent secure MIME formal should be utilized for KRR to KRA exchanges. Mr. Markowitz argued that such a specification was not required and would be overly limiting. Neither issue was immediately resolved. At the conclusion of the WG #5 report, there being no members of the public wishing to be recognized for the public comment period, the meeting was recessed for the day. Thursday, October 16, 1997 The meeting was reconvened at 9:00 a.m. on Wednesday with a continuation of the working group briefings. WG #8 Briefing The Working Group #8 (assurance/documentation) briefing was provided by Dr. Santosh Chokhani. He reviewed the table containing CKRS assurance levels, the assurance class, and proposed basic/enhanced/deluxe levels. It was not immediately clear to all members how all table entries were selected, and Dr. Chokhani suggested that the supporting documentation he would provide the Committee would help explain the options and his selections. After discussion, it was suggested that Table 2 would not be needed. The Committee may wish to consider having advisory statements added to the text instead. The Committee recessed for the remainder of the morning so as to allow working groups 1 and 5 and working groups 2 and 8 to meet jointly. The afternoon session was spent reviewing the results of the morning working group meetings. WG 2 and WG 8 will be making more changes to their submissions. WG #1 also will have some minor changes as well as WG #5. Therefore, it is anticipated that new versions of all reports will be sent out prior to the December meeting. It was suggested that these be sent out to the membership no later than December 8. WG #1 stated their hopes to have theirs out by the end of November. Chairman Kent requested that Committee members arrive in New Orleans having read all of the working group drafts with written comments on the drafts. This would allow for a detailed dialogue and critique at this meeting. He reminded them that there were going to be breakout rooms available Tuesday afternoon and suggested that the respective groups take advantage of this time and space to meet and discuss any comments on text submitted. It was requested that all presentation materials for the New Orleans meeting be sent to Ed Roback no later than December 8 so that copies can be made in advance of the meeting. WG #1 & WG #5 Status Report Mr. French reported that an action item list has been developed. Ms. Barker volunteered to be the editor of the glossary; a draft should be available for TAC review by November 25. All TAC members will be sent a copy by the November 7 for comments to be accomplished by November 14. A new draft of the Announcement Section will also be distributed, which may address the issue regarding signature keys. Some of the Interoperability sections will be reworded, some will become comments and become part of the overview and some will go to the Announcement Section, some will be incorporated into WG #2 sections. It was suggested that the WG #2 include a security warning about the implications for non-repudiation if the same key(s) is being used for confidentiality and signature purposes (which it was thought should be discouraged). The draft report from WG#1/5 is to be distributed to the TAC by November 25. WG #2 and WG #8 Status Report The two working groups have begun to rework the requirements based upon the three tiered approach. The TAC discussed the desirability and propriety of including (in a government standard) a "Tier 1" which might be appropriate for non-government use but would not be permissible for government use. NIST and the liaisons will deliberate on this. (ACTION - NIST/Liaisons) The WGs also hope to link to other NIST specifications (e.g., algorithms, FIPS 140-1). There was also a discussion as to whether the recovery information itself had to be recoverable. General Discussion Dr. Kent solicited a volunteer to develop a list of system level requirements so that component system requirements can then be reviewed to see if any system level requirements have not been addressed at the component level (or need to be addressed elsewhere). Any members wishing to volunteer were asked to contact Dr. Kent. (ACTION - Members) In order to help focus the TAC, WG chairs should publish, within the next 2 weeks (October 30), the action items needing resolution to all the members. The Secretary will distribute materials electronically regarding logistics for the December meeting in New Orleans. Dr. Kent thanked the members for their participation and contributions. The meeting was formally adjourned at 2:30 p.m.