[Moims-dai] My thoughts on the meeting with Vint Cerf

[Mike Kearney]

Well, I seem to find myself in the position of being Vint’s devil’s
advocate, so I guess I’ll continue.  Not to imply that Vint is the devil.  

 

*  In his discussion with us, Vint asserted that the success or failure of
OAIS hinges on the existence of successful implementations, 

Somewhat disagree.  His statements support he fact that OAIS is completely
successful.  Digital preservation fails without both OAIS and (in addition)
successful implementations.  

 

*  analogous to the IETF philosophy of requiring demonstrations of protocol
implementations prior to approval of RFCs as Internet standards.

He did reference IETF RFCs as a success story, but when he mentioned
demonstrations of protocol implementations, I thought he was referring to
CCSDS’s requirement for implementations (not only of protocols) prior to any
blue book approval.  Regardless of which organization he was referring to,
it applies to both, so there can’t be any issue that any implementation
(normative, testable) must be prototyped before approval as a standard.  

 

*  I would maintain that the specific implementations, while they might help
an organization to satisfy the requirements of ISO 16363, are not sufficient
to demonstrate a repository's ability to safely, stably, and indefinitely
maintain and preserve a body of digital data. 

Vint wouldn’t disagree with that.  Implementations aren’t sufficient without
OAIS management processes and the associated audits.  

 

*  Whether a warehouse uses Microsoft or Oracle software in pursuit of its
business goals is essentially irrelevant. 

No issues with that or your banking example.  To clarify, Vint wasn’t
suggesting that we standardize an implementation of an application.  His
advocated implementation standard is for whatever is required to access the
data, most likely with a different application.  Most likely many years
after the original application is no longer operable.  

 

*  However, for audit purposes, imposing a specific interoperability
protocol as a solution to some felt need for more reliable (more reliable
than what?) production of DIPs carries its own risks. 

Vint was OK with OAIS and OAIS audits.  He advocated an additional
access/implementation standard.  He didn’t advocate that the compliance with
that implementation standard should be audited.  Perhaps that’s still an
open question that he left for us to figure out after we have a concept for
that implementation standard.  But most organizations don’t use audits, they
use “gold standard” facilities to test protocols for compliance with a
normative implementation standard.  However, CCSDS is so poor that almost no
other CCSDS standards have such a test facility.  And as an aside, almost no
other CCSDS standards have an audit process, either.  But the point is

Vint was (in my opinion) perfectly OK with OAIS and OAIS audits.  He
suggested the implementation standard, but he didn’t suggest an
implementation standard audit.  

 

*   converting an internal to an external inferface for managing the
archived material would clearly open a window for deleterious accesses 

Within MOIMS, the SM&C group has a very layered architecture for their MO
Services.  For those services they use abstraction layers to convert an
internal to an external interface as part of their core philosophy.  Those
services (permit me to minimize DAI for a moment) are spacecraft telemetry
and command services, incredibly more critical (safety of human life) than
mere archives laying on the ground.  So if the problem of deleterious
accesses can be fixed for realtime spacecraft control, I’m sure it can be
fixed for archives.  

 

While on the MOIMS SM&C point, I wonder if the job of the implementation
standard should be turned over to MOIMS-SM&C instead of MOIMS-DAI?  I
believe that SM&C actually has an archive access service on their radar
screen as a future project.  

 

*  In closing, I do think that MOIMS-DAI should address the conceptual gap
between Interoperable Implementations and Safe, Reliable, Long Term Digital
Preservation. 

Right on!

 

*   I think that Vint and MOIMS-DAI see different horizons when we look
toward the future. I think his future is nearer than MOIMS-DAI's, and his
concerns are more imminent and urgent. 

Disagree, but I think I know where you got that impression...  The
“interoperability” discussion.  I think Vint’s interoperability suggestion
has two dimensions:  Interoperability between current organizations and
interoperability with future generations.  While talking to DAI, I felt he
was emphasizing the latter (future generations).  But if you could kill two
birds with one stone and have an elegant implementation standard for archive
access for present and future, that would be even better.  But at no time
did I hear Vint restrict his goal of interoperability only to current
organizations, excluding future generations.  

 

*  We want to make it possible for some Information Archaeologist in that
far future to be able to access, but more importantly, interpret the bits
that have been archived under OAIS conditions. 

Right on.  No conflict with Vint’s statements, right?   

 

*  But we must acknowledge that some will continue to implement long-term
repositories while emphasizing short term goals like interoperability and
fast access to DIPs.

With both dimensions, interoperability between current organizations and
interoperability with future generations, interoperability is not a short
term goal.  

 

Not sure if “fast access” implies something that I didn’t pick up on.  It’s
always better than slow access (for both current and future organizations).


 

So, do I sound like Vint?  Not nearly as smooth, I’m sure.  

 

   -=- Mike

 

Mike Kearney

Huntsville, Alabama, USA

 

From: moims-dai-bounces at mailman.ccsds.org
[mailto:moims-dai-bounces at mailman.ccsds.org] On Behalf Of Terry Longstreth
Sent: Wednesday, April 6, 2016 12:20 AM
To: MOIMS-Data Archive Ingestion <moims-dai at mailman.ccsds.org>
Subject: [Moims-dai] My thoughts on the meeting with Vint Cerf

 

This still isn't as good the original that disappeared from my new 'smart?'
phone.   For one thing, it's much longer than my original intent.
Regards,
Terry 

-- 


 

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/moims-dai/attachments/20160406/1aa99d64/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 11118 bytes
Desc: not available
URL: <http://mailman.ccsds.org/pipermail/moims-dai/attachments/20160406/1aa99d64/attachment.jpg>