CI Workshop notes 11/18/2002

CI-Workshop Organizing Committee Meeting
Monday November 18, 2002, 3 PM
Room 313 - Baltimore Convention Center

 Agenda


Notes thanks to Frances Van Scoy

Mel Ciment

Mel we're here through auspices of Centers Development Initiative
he used to be with NSF, retired 4 years ago from CISE
Rand Haley is project director

distributed handout, will send slides if needed
www.epscorfoundation.org/cdi

3 year project, funded by NSF EPSCoR program
EPSCoR Foundation + University of South Dakota
Joe Danek is PI
Mel Ciment is Senior Science Advisor
Anna Suarez, recently NSF, added recently
consulting and technical assistance to teams at EPSCoR universities competing for large scale support at NSF

impetus from conversations with Joe Thompson
re how to improve competitiveness of EPSCoR states

so far, assembling consulting teams to help people write proposals, do reviews

over $2M over 5 years to ITR they consider to be large scale

slide - map of EPSCoR states, including AK and NM, HI

why is it we don't get enough business in this area?
we want to look at cyber infrastructure for supporting large scale science and engineering

Rand organized an IGERT program this past summer
invited people from EPSCoR states
potential PIs
successful PIs from EPSCoR and nonEPSCoR states
program manager

possible future workshop on centers infusion
maybe we can find groups in EPSCoR states that aren't big enough to become Centers but could be attached to existing centers
workshop still on back burner

for this workshop
report on Cyber Infrastructure
deliberating on its draft since April 15
Kelvin is on that committee
Mel and others looked at draft and came to conclusion it was trying to fry some big fish
but not the small fish we're interested in
doesn't have an EPSCoR focus or strategy
might be buried in appendices

would it be appropriate to think on that level?

we got friendly responses from people in Bob's old division ACIR Advanced Computational and Infrastructure Research
Rich Hirsh very positive
Jim Hoehns very positive, EPSCoR office
CDI program had some left over funds

we want to
*survey/analyze the Cyber Infrastructure capabilities and plans in the EPSCoR states that broadly support the EPSCoR community's competitiveness in NSF Large-Scale Science and Engineering programs
evaluate effectiveness of major high performance computing and communications infrastructures
* provide recommendations on EPSCoR Cyber Infrastructure opportunities, requirements, barriers, and implementation strategies


political motivation =>

* recommend new areas of emphasis for NSF to support future needs of this community
* recommend a CDI technical assistance implementation plan associated with the above challenges and opportunities

bring around table people who are developing big proposals in the EPSCoR states

Bob B is in charge of this effort and has is formed this organizational committee

Mel's role - like NSF program manager
give you the money, you do the job, translate into action

Agenda

Discussion

John - what is next step?

Bob B - our goal is to organize a workshop
recommend directions
help people with proposals
suggest new programs to NSF
we have workshop charter on the website

asked Kelvin to give briefing on Atkins panel
when do you expect report to be out

Kelvin
target is by end of calendar year

Agenda

3:22 Kelvin

(slides)

(Cyber) Infrastructure

cyber infrastructure plays role in information economy that civil infrastructure did in industrial economy

The History of Planning at NSF (slide)
1985 launching of 5 original centers
Hayes report 1995 pivotal
our report picks up where Hayes report left out

idea of not just a couple of centers but an alliance to deliver much broader distribution of computing

our report goes much beyond this

Charge to the Blue Ribbon Panel

evaluate current PACI programs but not formally
no site visits
but how they might fit in the future

went far beyond recommending new areas of emphasis for CISE directorate

problem - how descriptive to be in implementation plan
new initiative is $1B, so business as usual at NSF will not suffice

draft released in April
http://worktools.si.umich.edu/workspaces/datkins/001.nsf

Margaret Wright at NYU, came from NSF
Stu Feldman is a VP at IBM
so good representation from academic and private sector communities

Information Sources (slide)

had about as many survey responses as did Hayes Report
about 700 responses
worded some questions to enable direct comparison with Hayes Report
are evaluated in appendix of blue ribbon report


our report WILL say something about EPSCoR
there were written transcripts produced from testimony

an all encompassing review
that is taking longer than anticipated

Some of the Findings (slide)

should be a director or leader of this initiative 
should be at level commensurate with AD
high enough to have visibility, authority, stature in community

trying to get other disciplines involved in cyber infrastructure
strongly reflected in surveys

as computers become more powerful they become more difficult to use
we're getting less and less of theoretical peak performance than 10 years ago

Some of the Findings (slide)

have not only responded to change but also have led change
but reflect environment of 1980s (batch oriented, not interrupt driven)
report balances extolling Supercomputing Centers and making recommendations (for how they should change)
problem due in large part to lack of funding

Some of the Findings (slide)

need a new IT professional, a computational scientist
expertise in scientific discipline, computer science, AND mathematics
bridge gap
know about basic computational issues

Knowledge Frontiers (slide)

report will talk about some things that are futuristic
but can point to examples where such things are currently happening

Blueprint (slide)

4 key issues

2nd
some is basic CS and IT resource
but some is developing services

3rd
these services are becoming expected for advanced research

4th
digital divide, haves and have nots
tribal colleges, hispanic colleges

budget slide
tried to be specific without being overly prescriptive

this is meant to be all new money
ongoing, incremental funds

Mel: to what extent did you look at what the basis is?

Kelvin: we looked at the base of funding that NSF has right now, the best we could understand it
what is being put into ANIR and ACIR and PACI

Mel: there is ITR money that cuts across
are you saying ALL of this is new money for applications?
that ITR money can disappear also

ITR is not sitting in CISE, it is sitting at some other level

Kelvin: idea is ITR will morph into something else

Kelvin: idea that this is new investment of $1B
not $600M new and $400M old

John: this report includes follow on money to ITR after ITR finishes

Kelvin: that is the assumption

John: I imagine you say that explicitly in your report

Kelvin: yes, I think so
we also mention KDI
you might re-create programs every 4-5 years within this initiative
there could be rolling initiatives within this particular billion dollars

Need Appropriate Organizational Structure (slide)

how to recommend to NSF that this should be implemented

leave it in CISE?

no, it has to be its own entity
has to cut across disciplines
not like polar programs sitting in director's office
but has to be in position that it has some recognition

a matrix management organization is what we're recommending

the way NSF is set up now, it is probably not able to handle a new injection of $1B into something like this

we walked fence between saying you need to create new office and saying outcome if you don't

Future of PACI within Initiative (slide)

recommend 2 year extension of PACI
otherwise PACIs would become unstable

then recompetition, new initiative

report we're working on now is substantially different from April draft

Agenda

Atkins Panel Discussion

John: you mentioned in passing EPSCoR
will there be something in report direction towards special needs of EPSCoR

Kelvin: yes
was one of the last things to be written and I haven't seen it yet
don't want to make EPSCoR states stand out as being "terrible"

John C
someone invented term I hate: geographically disadvantaged
EPSCoR states not near urban centers where cyber infrastructure grows up

Joe T:
a large number of the graduates are produced from EPSCoR states
one thing that is missed sometimes is that country is dependent on EPSCoR states to produce graduates, not just research
another very good reason to have EPSCoR states at same competitive level as others

Mel:
one of the really important slides is your blueprint slide
I imagine your slide has a lot of content behind it
was there any analysis done of role of EPSCoR along those dimensions
what the premiere efforts are going to look like

Kelvin:
I don't believe so
workshop on cyber infrastructure in atmospheric science in Boulder recently
… then EPSCoR states go and do their own analysis
we had people from EPSCoR states testifying
we didn't drill down that far

Mel:
I don't think we want to start over and redo the entire cyber infrastructure analysis
we want to start with this to the extent we agree it represents the EPSCoR needs
are you confident this reflects the structure of the final report?

Kelvin:
yes

Mel: having heard your discussion
our committee might look at your blueprint slide
and use it as outline of chores we might give ourselves
to parallel what you are doing

John:
the EPSCoR states have some very special needs since they're way out on the end of the food chain rather than where things get started

Mel:
if there is going to be a program that looks like this blueprint then we want to have an identifiable piece

John:
PACI has given lip service to EPSCoR and a few pennies
Dan and Barbara Kucera get some funding
co-funding works if PI is from EPSCoR state . . . . but needs to be changed

Kelvin: in creating this report, we realized cyber infrastructure is not very good
we had meetings by video conferencing and had problems

John: we were giving testimony and all Access Grid nodes worked except the one at NSF

Bob B: usually enemy of fiber is backhoe but here enemy is carpet knife

Agenda

3:30 - CI Workshop Charter and Discussion - Borchers & All

Dan, Mel, and I met about 3 weeks ago and set a date in April for the workshop (4/27 – 4/29)

Dan
Bonnie made comment that it is very near to Internet2 meeting which will be in Washington

Bonnie
not I2 but Net03, Educause

Dan
Net03 is Wed, Thur, Fri

Bob B
so could be synergy

Dan
Bonnie, Bob Aylward might attend

John Doyle might attend

April 27-29
reception Sunday night, adjourn at noon on Tuesday

Robert A
April 9-11 2003 is I2 meeting

Mel
we need more time

Bob B
we can't pull it off on that time scale
we’re being aggressive as it is

3:45 - Discussion of CI Workshop agenda, speakers and attendees

3:55

Bob
Who should attend?
shouldn't be entirely EPSCoR
who should talk?
Atkins to talk about The Report
probably no one at NSF could address probability of implementing
or getting funding

Mel and I were involved in 1993 report
Lou branscomb from IBM chaired it
was mostly ignored by Science Board

Hayes report was couched to Science Board as prelude to writing solicitation for PACI
Hayes went to Science Board in September
draft solicitation in October
for approval in December
fast action but no new money
hanging on to and reinvesting centers money

John
good analogy with Lax report
recommendations are practically identical
1982 recommended software, centers, just 1/10 as big
eventually in implementation plan recommended new organization with new money
which happened in 1984 was $100M

Bardon Curtis was implementation plan

Dan
they are ones who said something about networking

John
so NSF could put into 04 budget a new request as they did in 84

Joe
a lot of similarities to PITAC 99
that report said initiative ought to go beyond CISE

Mel
happily we don't have to worry about scope of issues Atkins report worries about
large implementation issues that they are mired in right now

John
wonderful opportunities

Bob B
by end of April won't have NSF response

John
they'll be in middle of preparing budget request

Mel
what are those things?
did they miss some things EPSCoR states need to do in training
make the focus that way
not worry about entirely paralleling everything in Atkins report

John
this is a very large effort for NSF and is infrastructureand if you let things go their usual way, infrastructure will all be built in California and Massachusetts

Bonnie
different perspective
not sure how many EPSCoR states are I2 members

Dan
all are

Bonnie
so some implied good connectivity
last mile is still an issue but some states are addressing
be more assertive and say not what we need but we can provide
we have some strengths that we can provide leadership on
what we have is critical mass
many of us have come together and made things happen

I thought of human capital
what can EPSCoR do to address human capital need?
curriculum?
computer science?
what crying need can we address by working together?

we have needs
we have strengths

Dan
as a stalking horse for what you're saying
one we thought about, at South Dakota they have EROS data center, huge archive
Atkins report mentions data archives
is FFRDC
has a lot of EPSCoR universities
Great Plains network - already has community

Bob B
where?

Dan
Sioux Falls, South Dakota
LANDSAT data for 30-40 years

Mel to Bonnie
did I get it right?
EPSCoR states because of peculiar lack of critical mass
partnerships and collaborations are critical to achievement in this area
is valid, very important issue
something that inevitably not going to get into report
because other states know how to partner

Joe
one thing you guys in Foundation might do
I'd like for somebody to get number of graduates in CS and computer engineering in EPSCoR states compared to entire country

Mel
you think it is proportionally much higher?

Joe
majority of African American graduates coming from EPSCoR states

Bonnie
we need to get away from saying that EPSCoR states are always in need
we have strengths we can offer
we have a very strong cs program accredited through Ph.D. level but is very small
doesn't have critical mass and isn't recognized

Mel
one of the important things is to attract outstanding people to workshop to show them
when we hold up this document, it should be multi-use document
PR
build program
explain our needs
demonstrate we have capable people

Bonnie
but don't want workshop where we talk about how good we are
but come out with action plan

Joe
if had stronger educational partnership among these states
one of our resources is a graduate program MS and Ph.D. in computational engineering
because of ERC but is very popular program now
we could have students come from other EPSCoR states to spend part of graduate program at Mississippi State 
partnerships could make some very important things happen

I submitted an unsuccessful IGERT proposal in mid 1990s to do that in partnership with NCSA and Rice
reviewers said, who does Mississippi State think they are to think they can play with the giants

John
sent similar proposal from Kentucky with similar reviews

Bonnie
we might still have some infrastructure issues
goal to contribute more to human capital
may need more robust video conferencing or other tools
we could be a model for some kind of human development program
rather than always being recipients of infrastructure
we have enough that we can start feeding into the system

Bob
are you suggesting an emphasis in workshop and report on human capital ebbs and flows

Bonnie
suggesting we get away from saying we always need more infrastructure
but get to saying what we can contribute

John
the two can be combined
we need the infrastructure so we can contribute

Dan
leverage that human talent

John
vast majority of students in country go to school within 50 miles of home town
a lot are very talented, need place to go that is connected to resources rest of country has

Bonnie
what would we need to know or do in order to make a very successful conference to get good outcomes to put into food chain

Mel
ideal world is never realized
study every university, needs, where they are going
Atkins report couldn't do it
they had advantage of looking at program that has existed 15-20 years
we need people able to look at own needs and project out to community needs
could go along many dimensions of blueprint

Joe
important to look at things we can do for selves with some help
help some universities in EPSCoR states sell their admin on importance of infrastructure
a few years ago I advised a university on ERC proposal; they did not submit because they could not muster institutional support
some of us could help with success stories

Bob
similar problems getting Hawaii interested in ITR proposal when realize probability of success is not high

Kelvin
found section in blue ribbon panel report on EPSCoR
1.5 page title underrepresented groups in EPSCoR states
last paragraph
initiative should embrace EPSCoR and … activities toward EPSCoR states . . . . should participate in all 4 . . . model could be applied to MSI (Minority serving institutions)

Dan
asked Kelvin to mail it to committee

Bob
ci-workshop@bborchers.com
will go to everyone including consultants

Bonnie
word targeted bothers me
sounds like making exception to include
I'd like to be included

John C
there is program in Europe, run by the EU
said if you set up an international consortium, you should include some of the less developed countries (and get higher priority for funding)

Dan
STC does that

John
I don't think they pay attention

Bob B
not enough points

Mel
2 things
(1) respond to current draft Kelvin is working on
(2) keep our slate open

Bob
might be useful framework as outline

Mel
is what I said earlier
we should take our energy and decide how to organize our effort
o.w. we could take a day looking at their report

Bob B
there are EPSCoR programs in other agencies
do we want to include people from other agencies?

Mel
only so far as they might inform on Cyber Infrastructure
NIH has enormously large EPSCoR program, IDEA $125M
CObrE, brIN
but what we're doing here is an NSF project

forefront of this: our funding comes from NSF
to the extent that large scale activities are promoted by advanced cyber infrastructure
some institutions can't make that
cyber infrastructure needed for whole host of competitiveness 
not just Atkins mandate
I don't want to restrict you but core activity
how does cyber infrastructure support competitiveness

what they spend on Atkins report? over $1 million
we'll spend $30-60K

role of EPSCoR universities in producing students in various areas
partnerships in EPSCoR states

John C
nature of cyber infrastructure is that it is distributed
but big NSF programs connect network in California to network in Illinois
TeraGrid

Bob
less and less true
distributed biological observation centers
seismic stuff
are much more distributed

Mel
John is right
I asked George Strawn, he was spokesperson at the time
now moved up to CIO position
his answer was
EPSCoR states were invited to use TeraGrid facility that was being created at 6 institutions
they did not have a plan on how EPSCoR universities could help create

Bob
TeraGrid was response to broader cyber infrastructure solicitation
TeraGrid had enormous advantage - people already heavy players

thinking 50 people at workshop

John
really have to push hard to get people from EPSCoR states to come

Bob
True, but can't be just EPSCoR states talking to each other

John
need leaders, people building cyber infrastructure

Mel
Peter Freeman and others are supportive of workshop
will get people they support to participate

Bob
Peter is showing up here in about 45 minutes and I plan to talk to him

Dan
number of people is still . . . .

Mel
look at blueprint areas
try to figure out how many you want in each area, who will take lead in each area
I don't have a good feeling of who are leaders in EPSCoR community we need to get to workshop

Dan
I could go around this room and ask

Mel
outside leaders are much easier to figure out
I want to make sure, one reason this group is very important is to make certain we have right people from EPSCoR states

Joe
what is desired result of workshop?

Mel
I gave you several recommendations for people to look at , take to institutions, lead to collaborations
bottom line from us--we want to see funded proposals in large scale categories

Joe
might be recommendations to EPSCoR states, to NSF itself

Mel
I don't shrink from opportunity to go back to Atkins report that Borcherds report come in and . . . 

John
could pick up examples Bob mentioned
GriPhyN, NEES
and role of EPSCoR states

Bob
could nucleate EPSCoR states
balance various strengths

Bonnie
tout it as an EPSCoR response to Atkins report?

Mel
no
is one of several activities

Bob
Atkins report will precipitate interest in cyber infrastructure

John
you want to look at opportunities to broker partnerships
even more important between EPSCoR and non EPSCoR states
most productive partnership I brokered at Kentucky was with Illinois

Mel
which activities go which way
this culture that you have, it's not clear that there is just one way

George
EPSCoR states geographically disadvantaged
ex. Maine near Vermont, other near states are non EPSCoR

Mel
new areas of emphasis
recommendations to NSF for areas
and new mechanisms for supporting EPSCoR
what kind of program should be out there
how to get at these issues
what did we do with ITR?

Dan
new mechanisms for support
one of few programs that allowed for several states to get together was EPSCoR Grants
last one was Access Grid that Barbara wrote
SEPSCoR was one
more than one school in more than one state got funded

Bob
Access Grid is important infrastructure

George
we built our own

Dan
is difficult for collaboration when go multistate

Bonnie
House and Senate have passed bill to provide lots of money for cyber infrastructure security
NSF and NIST to get a lot of this money
is anyone watching to see where that will end up at NSF?

Mel
is excellent

Bonnie
from curriculum standpoint we are watching

John
good question to ask Kelvin
do you expect to see something in Atkins report on cyber security

Kelvin
I don't recall if we mentioned anything specifically on cyber security
I think we did mention issues of security of data, but national security, I don't recall

Bob
this wasn't on event horizon when panel was chartered
Mel put together a series of panels on Friday on homeland security

Mel
Cyber Security Act, Chairman Vollard pushing
has passed both houses, will probably be signed
$900M+, is a lot of money
I go two ways
(1`) it is really important for EPSCoR states to get involved
(2) CDI may not get a lot of business

we want 
this project is not just doing
this workshop is not to do all the needs for cyber infrastructure for EPSCoR states
too vast a subject
focus on high points of support for large scale science and engineering
cyber security will come out of this

Bonnie
document out of White House has higher education section
we should be come familiar with this

Mel
bring in people who can talk to this from federal side, research side, EPSCoR side
important to get structure of what we think we can do in a day and a half
try to get people who can drive this

Bob
what is next step?
do we make some homework assignments
do we get back together by teleconference?

Mel
all of the above

Kelvin
we had this cyber infrastructure of environmental research education 
web site
planning documents
who do we invite
what are topics
what are outcomes
could be useful in terms of people we had speak
Kim Mitch from Oklahoma talked

great opportunity if someone got document I sent if you can give me some feedback on EPSCoR components, especially last paragraph

Bob
assignment for individuals
is short fuse before we meet again


Dan 
individually we can send to Kelvin

Bob 
send it to reflector as well

Kelvin
thought you could caucus

Joe
PITAC99 and EPSCoR
this has been noted 3 years ago and we still haven't addressed the problem
NCO.gov website

Kelvin
can you send it?

Bob will send it to Kelvin

Bonnie
Educause has security web site
educause.edu/security
series of workshops

Dan
aren't those mostly for network administrators?

Bonnie
no
having for research administrators

Mel
assuming we have right people at table for workshop
we should ask each person here to submit list of people, speakers, invitees (not the obvious ones from NSF and world known)
university, government, industry types in EPSCoR states

Dan
CDI data base of recent awards over $250K
PIs

Mel
speakers
invitees

Bob
will send out email reminder
we have clearer picture
but likely won't get further this afternoon

Joe
appropriate to invite some congressional staffers?

Bob
of course!
Inouye is highly defense oriented
Ted Stephens is chair of defense appropriations and Appropriations

John
about half of appropriations chairs are from EPSCoR states now

George
a goal of workshop is to generate more proposals
so helpful to invite people who have projects who need partners

Mel
could have people from EPSCoR or non EPSCoR states who have been successful in building centers
what did it take to be successful
is what we did in IGERT case
would like to look forward; what are requirements
who is looking broadly across needs

4:30 - Adjourn

Bob
I like a face to face meeting so I know who I'm dealing with
I'm going back to the island on December 3

do we all have Access Grids

Mel
I don't

Dan
come over and use my PIG

Bob
anyone with a machine on multicast enabled network can bring up pig

Mel
groove.net
allows you to set up secure file sharing on desktop environment
people can go in and not access your other files
can be helpful when we get to point of creating documents
but for now, teleconference at a specific time

Bob standing staff meeting at 3pm on Wednesdays

John
1st week of December is allocation board meeting

Mel
try to get on every two week schedule

Bob
will send out tentative dates and times

Mel
every two weeks
prefers phone

Bob
can bridge phone to Grid
Hawaii is 5 hour time difference
December 9 and 10
try 3pm Monday December 9

4:51 pm adjourn

Agenda

==============================

Return to top page