Re: [dsii] Potential IETF Work Items

[Scott Brim <[e-mail]>]

On Tuesday, August 14, 2012, Russ Housley wrote:

> At the DSII BOF in Vancouver, Andy Maffei gave a list of potential work
> items.  Is there interest in pursuing any of them in the IETF?
>
> Tom my memory, Andy raised this list of potential action items:
>
> * Summarize existing schemes being used for PIDs
>
> * Develop recommendations on a limiting number of PID schemes
>
> * Define an "Interoperability API" for registering dataset identifiers
> (perhaps based on the EZID API)
>
> * Specify a core set of metadata to be provided by all dataset (probably
> includes a number of dublin core attributes as well as a digital
> fingerprint)
>
> * Specify a set of conventions for resolving PIDs associated with datasets.
>
> In addition, there seems to be some overlap with the CDNI WG.  They are
> also considering some metadata issues; see, for example,
> http://datatracker.ietf.org/doc/draft-cjlmw-cdni-metadata.
>
> Should the IETF specify conventions for paths of PIDs for data objects?
>  These might include location metadata, access policy, authentication
> methods, authentication credential, and so on.
>
> Russ
>

These issues are very significant to us (Internet2), as we serve
researchers and they ALL have data issues.

The problems are large but I also wonder what part of this is in IETF
scope.  I asked about this in the halls after the BOF.  If this work takes
place in the IETF I will be very involved, but IETF expertise is primarily
in protocol engineering, and I don't see much need for that here.  There is
a slight overlap with CDNI but not much.  However, we (Internet2) need to
work on these issues somewhere, and if the IETF is glad to host the work,
we're glad to be active in it.

-> I suggest a side discussion at the October DSII meeting in Arlington
about specifics of how the IETF could help.  (Alas, I can't be there, it
overlaps with our member meeting.)

Regardless of where the work is done, I do not think we should attempt
"limiting number of PID schemes" (they, like collaboration environments or
mail apps, are very subjective and attempts to quash inventiveness will be
futile).  Instead of putting large amounts of effort into unification and
not completely succeeding, we should try for federation: define interfaces
for interworking while also supporting local innovation.   The "core set of
metadata" is that which is to be presented to the "interoperability API"
... whether metadata is constructed that way internally should be an
independent matter.

Scott