date: Wed, 14 Feb 2007 17:26:22 -0500
from: "Bruce A. Wielicki" <b.a.wielickiatXYZxyza.gov>
subject: Re: Teleconference summary
to: Ferris Webster <ferrisatXYZxyz.udel.edu>
Ferris et al:
Sorry I missed this telecon but I read through the minutes. I agree with the summary but
with one serious concern: the document attached summarized:
At the moment, CLASS is still focused on large arrays. It must evolve into an enterprise
solution. To achieve that, an enterprise statement of requirements must be developed,
rather than simply organizational statements of requirements.
Since the original motivation or CLASS was making NOAA able to handle new large data
storage commitments like NPOESS, I am puzzled on why CLASS should be expanded to an
enterprise solution for all NOAA data sets. We clearly saw that the diversity of NOAA
types of data is shockingly diverse. Developing a one-size-fits-all "enterprise" solution
seems like the wrong way to go. It is likely to be very expensive, very complex, very
inflexible, and will please no one.
Since the original congressional logic was dealing with new large data volumes: it seems to
me that the focus of CLASS on large data sets makes perfect sense. As the "Death March"
book reminds us: 80% of the functionality comes from 20% of the requirements. Pick the big
new problem (NPOESS) and do it right. It won't be well suited to fish guts data or to
individual surface site temperature records: and thats a GOOD thing.
From the outside it looks like CLASS is way underfunded in NOAA to even do that one thing:
large satellite data sets. If that was the original motivation, and that is the tallest
pole, and funding is very tight: there are only two things to do: spend all the money on a
glorious architecture that will never be built (death march), or triage the requirements
down to the really critical thing that must be done. If CLASS is successful, build on and
extend it. I think congress is leading noaa down the death march path of requiring way to
much for way too little time or resources. For any of you that have not read this book: I
highly recommend it (an easy find on amazon and a used paperback copy is cheap). Lots of
that rare common sense stuff. After living through EOS, EOSDIS, and CERES with ~ a million
lines of code, I can confirm his conclusions.
Given that CLASS is trying first to do new large NPOESS satellite data sets: I'm rather
to hear nothing about using lessons learned from the NASA EOS satellite system which is
flying instruments with very similar data rates, volumes, data products, and users (from
science to commercial). The Terra mission alone in 2006 delivered about 12 million data
files to 25,000 unique users. The CERES data products on Terra alone in 2006 shipped out
40,000 Gbytes of data products to users.
The IPCC policymaker summary has just come out. The full report is later this spring. The
NRC Decadal Study has just come out. Climate change and the global data sets needed to
support it are going to increase in importance. This panel should be comfortable that
CLASS will be ready to catch the huge ball called NPOESS and be able to serve it up in nice
sized chunks to the user community. This should be priority 1 for this panel to verify as
something on track. I don't see it yet. I'm hoping I just missed it because I missed the
telecon. Is that it?
At 3:58 PM -0500 2/14/07, Ferris Webster wrote:
Attached is a summary of this week's DAARWG teleconference. If any of you who attended
feel that I have missed something or have misinterpreted your statements, please let me
know and I will modify it.
DAARWG members, please note that the document sent out on Monday, the GEO-IDE Concept of
Operations, is a key document, that summarizes the approach being taken. It should have
been circulated earlier, but there was (understandable) confusion about which documents
to send us. I urge you to take a look at it.
Content-type: application/msword; name="0702 notes.doc"
Content-disposition: attachment; filename="0702 notes.doc"
Content-description: 0702 notes.doc
Attachment converted: Macintosh HD:0702 notes.doc (WDBN/�IC�) (001AE4A5)
Bruce A. Wielicki
Mail Stop 420
NASA Langley Research Center
Hampton, VA 23681-2199
Phone: (757) 864-5683
FAX: (757) 864-7996