Long note, only for folks interested in programmatic design issues for
interoperability, re muscle work. Don't read if your muscle interests are
about code, and/or installation practices, etc. There is no code, here!

----------------

I invested two hours reading the goal/motives/introduction sections of the
US government's latest reporting level contribution to the smartcard and
id-card efforts in the US government employment and contracting world (which
is also a contribution to a set of international norm setting processes,
also, evidently). The main objective of the analytical reading was to see
how the world is evolving around our movement, and thus how the open source
design world can take account of these ongoings, if at all.

http://csrc.nist.gov/publications/nistpubs/800-73/SP800-73-Final.pdf.

The acknowledgements to the SP note Scott's role in authoring the cited
document, and note the special contribution of Booz Allen Hamilton for
technical and editorial contributions.  I will attempt to discover who paid
for BAH's work product, from public records; for all we know, it could be
BAH itself, making an admirable, public contribution. AS we should all note,
this is an SP, not a FIPS: it's a report --- on "good stuff", related to
standards work, - but not a standards effort in itself.

(a) some of the early goals language will be clearly unacceptable to ISO.
ISO normative standards don't "govern", where NIST standards and their SP
support document under the modern FISMA do now "govern" US folk. As always
in the US-speak used in its executive branch FIPS and related documents,
normal words have very technical and legal meanings. FISMA changed NIST's
authorities to now include the right "to govern" public security, recall. At
the same time, the intro to this _report_ only cites OMB A-130, as the legal
basis for the process of its issuing. How the SP contents relates to the
authority in FIPS 201 is not clear; the language usage suggests the document
production programs are related (as you'd expect, its one real team after
all) but somewhat confused in their authorities.

(b) "FIPS 201 also specifies that the identity credentials must be stored
on a smart card." While this may seem to mean what the world thinks the term
"smart card" means, we NOW  know the authors (of the SP) intend the term to
means anything acting like a smart card - including soft smartcard "PGP key
ring". If I understand Scott, its intended that your (enhanced) PGP keyring
in software could be a "smart card", subject to the governing usage controls
of PIV, assuming that said key ring satisfied the PIV interface
requirements. I.E. politically, PIV as expressed in the SP is _intended_ to
embrace and to govern all forms of cryptomodule, assuming it can emulate a
7816-4 supporting device that is PIV compliant. 

(c) What is interesting is that the interfacing requirements also include
ICC (i.e. chip level) interfacing requirements. That is, only certain chips
of certain functional parameters and assurance levels can be PIV compliant,
in the wider sense of the PIV II initiative.

If the PIV is on a soft crypto module, on the PC, presumably, the Pentium
has to play the assurance role of the "ICC", and be "PIV conforming". More
likely, the TPM controlchip is intended to play this role, as Scott hinted.
This means the TPM is being envisioned as going far beyond its current role
in trusted computing - securing OS integrity, and enforcing US crypto export
rules for application of keying material, into being a(nother) repository
for personal id. 

Presumably the right to boot a PC, or a VM within the PC or over RDP, and
the right to use crypto (locally or remotely in the hosted VM), will be
functions that are will only be armed when the PIV fragment of the
motherboard's TPM has verified the PIV-borne credentials. That! Or: all PIV
capable smartcard will be themselves TPMs, in 7816-1 module form. These
forms of TPMS are now appearing from in major chip making companies. Scott
indicated that PIV devices are supposed independent of the "media", by
design intent, and can be TPM hosted or not TPM hosted. The SP should be
read in that light.

(d) All the US agency specific stuff needs to get removed, before the
document can be referred to, in an ISO setting. The wider world doesn't care
a jot about OMB (a White House authority), and its stick and carrot role in
rule-ing over and then auditing other US executive agencies and depts' use
of their White House budget allocations. Now that we know about these
disclosures, however, the issue of OMB authority and influence is relevant,
in adjudging the PIV initiative, in the wider ISO sense. It is relevant to
note that, today, we can view OMB these days as the legal enforcement arm of
a more politically involved NIST, and its security programs under the public
law. One has to be aware of these relationship though, internationally,
amongst the US stakeholders who are influencing the PIV document content.
These are issue relevant to analyzing the US contribution to ISO.

(e) Looking a bit forward in the document to the ICC related sections, we
see mandates on Crypto algorithm support. This will need to go, in the ISO
format. ISO doesn't mandate crypto requirement support, ever. We reference
the ISO registries of crypto mechanisms, schemes, etc. National govts choose
their crypto. Period. Thus its for the a multi-lateral operational profile
of PIV, not PIV in ISO form, to agree on the common set of algorithms and
their parms, so we can wander with our ePassports anywhere, effectively. 

More widely, US power and its control over crypto algorithm access and
distribution have to be addressed in the politics of the ISO WG and plenary
processes, as do North American patents on ECC algorithm varieties. Etc. We
cannot have ISO used as means of projecting Intellectual Properties policies
that suit the US economy, or its particular export control objectives, or
its own national security goals re crypto strength, key lengths etc. All
these factors are PIV related, and are part of the wider ISO process. At the
same time, working compatibility with the features of the world's largest
economy has obvious user benefits...

-------

That's enough broad review on the program itself, I think, given publication
of the SP - which we should thank the US govt for releasing so the US public
can see a little of what's likely to be going on, in ISO, on their behalf.
Its always fun to watch government in action on crypto-related matters -
even if we have to do it by proxy via the analyzing the public SPs and
FIPSs.

Peter.
_______________________________________________
Muscle mailing list
[email protected]
http://lists.drizzle.com/mailman/listinfo/muscle

Reply via email to