>>Welcome to the
PTI seminar for May. I’m very pleased that we
are going to have a reprise of the People’s Memorial lecture
which took place last month at the Herman B. Wells Library
and also by Dr. Craig Stewart. And, Craig has been my
mentor for a very long time. I’m very pleased that we’re
able to have the talk here again at the cyberinfrastructure
building to let a lot of the UITS employees who might
not have wanted to make the trek over to the library that
day, so they have a chance to hear this really great talk. Craig’s going to take a look
back at where IU has been in computing and
cyberinfrastructure, assess the value of IU’s
current implementation of information technology and
support of research scholarship and creative activity
and suggest strategies for supporting IU’s excellence in these areas in
the coming years. And, he’s currently
the Executive Director of the Pervasive
Technology Institute. Dr. Stewart, take it away.>>Robert, thank you so much. [ Applause ] And, ladies and gentlemen,
thank you. Let me tear right into the talk and actually the first
few slides I’m going to click right past. We have a diverse audience
today from 12 years old to considerably more
than 12 years old and a diversity of backgrounds. So, I’ve got handouts
that went out by email with the announcement and
which are on your chairs which have a number
of definitions. And, I’m not going to
read definitions to you. You can look them
up if you want, and if you don’t need
to, you don’t have to. But, what I’m going to
talk about today is sort of where we started at
Indiana University in terms of computing, in support
of research, scholarship, and creative activities. I’m going to talk
about two decades of really steady progress
in service capability and service delivery from what
was originally called research in academic computing and is now
called research technologies. I’m going to do a little bit
of a rewind and recap some of the history and current
structure and function of the Pervasive
Technology Institute. Then, I’m going to do a little
bit more of a rewind and talk about where we are in terms
of national strategies and information technology
in advanced computing because that plays a
lot into where we are in our currently environment
locally as well as nationally. And then, I’m going to talk
about how we move forward, and I want to say a little
bit about purposes at hand and make a couple disclaimers. So, this talk focuses on
decades’ worth of work by hundreds of people. Most of what I’m going
to talk about was done by somebody other than me. I am particularly focused on
what research technologies and the Pervasive Technology
Institute have done, often in collaboration with
the School of Informatics, Computing, and Engineering. I’m not going to talk about
things that are relevant to networking and research
in creative activity. Those weren’t my portfolio. There weren’t things
that I had to do with, and you could do another
really, really long talk about all the cool things
that have happened as a result of IU’s involvement
in networking. And, last but not least, I
want to mention that the groups that I’m going to
talk about were formed in their current form largely
by Christopher Peebles and led by him for many, many years. So, with that, I’m going
to tear through four slides of definitions and
not read them to you. I will come back to
one of them later. You’ve got copies of the definitions you
can read if you want. What I really want to
do is start talking about where research computing
began at Indiana University, and it began in 1949 with
a grant from the Office of Naval Research to
Indiana University for a mechanical card-based
calculator which turned out to be so important to
chemistry and physics research at Indiana University
that in 1953, then President Herman B.
Wells called for the creating of a research computing center for Indiana University
generally. That center was formed
forthwith, and in 1955, astronomy professor Marshall
Rubel was named the first Director of Indiana University’s
Research Computing Center. And, he was not just
an eminent astronomer. He had a real sense for the
importance of computing, and one of the things that
he did as the first Director of the Research Computing Center
at IU was establish a principle that we have lived by every
day from 1955 to today. That is that the university’s
computing resources are a common good. They’re intended for use by the
entire university community, and in particular,
Professor Rubel’s view was that IU’s electronic computer
was as important and should be as accessible to his astronomy
graduate students as it was to his faculty colleagues
and the business school. He went on to write a paper in 1960 called the
electronic computer, as astronomical instrument that really revolutionized
observational astronomy at the time. Sadly, he died on
sabbatical in 1968, but he really put research
computing at IU on a good, solid footing and clearly bears
credit for where we are today in terms of our computing
environment and in terms of our physical environment
because we are in the Rubel Commons in
Cyberinfrastructure building. From that auspicious start,
there were intermittent times of a real distinction
of Indiana University. In 1964, two professors
of computer science and one staff member in
the computing center, Steve Young who in some ways
is my organizational great grandfather, wrote a
compiler called FASTRAN, FAST fortRAN compiler for
the IBM 709 computers. That was one of the first
ten software programs every protected by copyright. It was as a compiler, faster
than any compiler ever made by IBM for the 709
computing series, and the two computer
scientists involved in that project credited
Steve Young in his knowledge of the IO system, the 709 for
why the compiler was that fast. Going on, 1968. Computer science was formed
as a distinct department in the College of
Arts and Sciences. In 1979, Doug Hofstadter
published “Godel, Escher, Bach”. This is a photograph
of my soft cover copy of “Godel, Escher, Bach”. It was an amazing book that
revolutionized a lot of thinking about information technology,
information science, computing math, and
music as well. In the 1980s, many,
many people believed that Ernie Davidson
might be on his way to winning a Nobel
Prize in Chemistry. He developed the Hondo series
of analytic chemistry software. IU invested a lot of
computing resources in his lab. Unfortunately, another
group got a Nobel prize for work very closely related to what Professor
Davidson was doing. So, he did not get
a Nobel prize. 1993, IU acquired
an Intel paradigm, IU’s first parallel
supercomputer for general purpose access
throughout the university. IU had a couple other
parallel computers in individual departments,
but the paradigm which was a brilliant piece of parallel computing
engineering was the first university-wide accessible
computing system run by the Computing Center for
the general IU community. 1995 and 1996 two really
historic accomplishments by Dennis Gannon. First, he received a shared
university research grant from IBM to get an
IBM SP supercomputer for Indiana University,
and he participated in something called the
I-Way which, for one week at one conference
demonstrated what could be done with a ten gigabit
per second seashore to seashore network link tying
together multiple supercomputers in the U.S. that really was
the beginning of what we think of today as modern
computing grids. And then, you know, the ’80s
and ’90s, research computing at IU wasn’t distinctive
for very much else other than for contributing to
a revolving door effect in which bright, young
professors of physics, astronomy, and chemistry
would come to IU. They’d get tenure, and then
they would get recruited away to go someplace else with
better computing resources than Indiana University had. Then, everything changed
with Miles Brand as President of the University when he
recruited Michael McRobbie to come to IU as its
first full Vice President in information technology. And, Miles Brand himself
was not particularly and acutely detailed
in his understanding of information technology, but he understood the
strategic value of IT from the university standpoint. So, he set this goal for IU to
be a leader in absolute term in use and application of information technology
among American universities. This was such a ridiculous
goal at the time it was set that people working for the
Computing Center couldn’t suppress their giggles
at meetings. It was just outlandish to
think of IU being that way. And, yet, in 1997, Vice President McRobbie
led a reorganization of the IT organizations at IU to create what is now University
Information Technology Services. We created an information
technology strategic plan for the university. Copy of the cover
is right there, and what was really
important about that plan was that it was an IT Plan for the
University, not a strategic plan for the IT organization. It covered the entire
university, all of its activities and
enterprises, and the summary of this plan is get the
technology stats right. No busy signals on modems. Get networks right. Get administrative systems
right and invest heavily in research computing. And, in 1996 and 1997, President
McRobbie got engaged in a grant that was just in the
process of being awarded to Indiana University called
SCAAMP, Scalable Array of something something
Multiprocessors. And, the original grant
award from the NSF was going to buy a supercomputer and
one visualization environment. And, Michael McRobbie said, “No. We’re going to add some money
to that, and we’re going to make this really important.” So, what we got out of SCAAMP
was a 64 processor Origin2000, the first Origin2000 owned
by any individual university in the U.S. There were
other universities that had Origin2000s as
part of federal facilities, but we were the first
university to own our own. And then, we built this
thing called a cave which is a ten foot
cube, immersive, stereoscopic visualization
environment that makes you feel like you’re standing
inside a virtual world. And, we got this thing called
an Immersadesk that was put in Indianapolis which gives
you the same immersive, stereoscopic effect, not
quite as profound as a CAVE, but it’s pretty darn good. And, this was the very, very
first facility of research and academic computing that was
installed on the IUPY campus. And, these things
were great hits. You know, if, if a picture
is worth 1000 words, then, you know, a 3D visualization
is worth a lot more than that. And, in fact, when Mikhail
Gorbachev visited Indiana University in 1997, he
took a tour of the CAVE. Eric Wernert and
Eric’s colleagues dolled up a 3D visualization of the
solar system to include Mir. And, I wasn’t at that
meeting, but reportedly, former Premier Gorbachev
looked at the visualization and said, “Ah, Mirski.” So, this was the
beginning of visualization and computing together being
really, really important, both in supporting
university research and making the university
community aware of what we were doing. And then, President, Vice
President McRobbie worked with Lilly Endowment to
get what was the first of now many large grant awards
from the Lilly Endowment at Indiana University for
something called IPCRES, the Indiana Pervasive
Computing Research Initiative. $30 million, half of which
went to fund up the new school of informatics, half
of which went to create the pervasive
technology labs. And, this established
the three legged stool of research development
and delivery that has been the foundation
of our support for research, scholarship, and creative
activity at IU ever since. Basic research in what is now
the School of Informatics, Computing, and Engineering
applied r and d, very often taking new
innovations from SICE and converting them into
tools that are usable by every day researchers in the
Pervasive Technology Institute. And then, services systems,
consulting, resources, computer storage
systems, food systems, delivered by research
technologies. And, there are those
three organizations. Presumably, many
people know of one of the two recreational
uses of nitrous oxide. There is a second which is to make cars go really,
really fast. This is the go baby go button
which is commonly installed on the shifter of cars with a
nitrous oxide system in them. And, when you press that button,
the car goes really fast. I happen to have my
own go baby go button. [ Laughter ] And so, I’m going to rip through
a whole bunch of accomplishments that research technologies
and PTL and PTI have enables at Indiana University. So, hang on to your seats. Here we go. 1999, we installed a tape robot
in Bloomington with some very, very secure software
called HPSS, designed to make sure the
IU could preserve the data and materials being produced
by the scholars and researchers and artists of Indiana
University. We, in 2000, negotiated the
first enterprise site license for SPSS for any university in
the U.S. For many researchers at IU, the SPSS site license
is the most important thing that [inaudible] ever did
because it made it possible for any researcher
at Indiana University to have the social
package, statistical package for the social sciences
software. We got a major award from Lilly
for something called INGEN, the Indiana Genomics Initiative. $105 million, $6.7
million of that went into research technologies
to improve out ability to support biomedical research. And, a couple digital
libraries things. So, research technologies
was very involved in the digital library
project with libraries, and on the left is
a, is an example from the US Steel Gary
Works Photo Collection. Here is a picture from the
Cushman Photo Collection. We also, in 2001, helped put
online the alchemical notebooks of Isaac Newton which really
forced the physics community to reassess its own history. They proved conclusively
that Isaac Newton, the great empirical scientist,
the father of modern physics, spent most of the last years
of his life diligently trying to turn led into gold. When I was a physics student,
I was, when I was a student as an undergraduate in physics
classes, I was told, “Yeah, you know, Newton messed
around with that a little bit, but he wasn’t really serious.” You read his notebooks
online, and it’s clear that it was the major part of
his thinking for the last years of his life, and I think that putting those
notebooks online may be one of the most important things
that Indiana University has done as a contribution to the study
of the history of science. 2001 was a big year. We added another tape
robot in Indianapolis, and research technologies
created software such that people could
write data simultaneously to tape robots in
Indianapolis and Bloomington. So, if a tornado destroyed
Bloomington or the earth opened up in a giant earthquake
through the New Madrid Fault and swallowed Indianapolis, as long as those two events
didn’t happen on the same day, then the data resources
stored by Indiana University and what we call a scholarly
data archive would be preserved. This is one of many
examples of UITS and Research Technologies
taking advantage of the two campus structure
of Indiana University to mirror data storage
and computing facilities between Bloomington
and Indianapolis. We upgraded our IBM
SP supercomputer to be the first supercomputer
in the U.S. owned by an individual
university capable of more than a trillion mathematical
operations per second. We received $1.8 million in
the National Science Foundation to create something called
AVIDD, Advanced Visualization for Instrument Driven Data which
put vis and computing facilities at IU Northwest, IUPUI,
and IU Bloomington and really advanced
the capabilities and computing services
available for people that did data intensive science. The NSF funded IU’s
participation in something called the
TeraGrid, a national grid of supercomputers for use by
the national research community. This was the first time that
we were really beyond our own campus. This was the first time
that we had federal money to serve the national
research community. And, oh, yeah. And Eric Wernert and
John Huffmann, the elder, the chemistry professor, and
John Huffman the younger, the IT professional, created
something called the John-E-Box, this one liter cube which for
a few thousand dollars allowed anyone to have a stereoscopic
visualization environment in their lab. One of many things
that AVL has done to make high quality
visualization accessible throughout the university. And, here on the right, there’s
just an example, just a diagram. Bloomington, Indianapolis, IU
Northwest, systems redundant between both campuses
[inaudible] doing both campuses so that we could
keep our systems up and running all the time and preserve IU’s data
reliably no matter what. From 2003 to 2005,
there was a real shift in the national strategies
for advanced computing. Before 2003, the National
Science Foundation, Department of Energy, the
other federal funding agencies, thought of supercomputers as
if they were time machines. The function of the
supercomputer was to take a person doing a
very, very important problem and take them into the future
of computing capabilities and let them do something that
would be generally possible for other researchers
only years in the future. In 2003, The National Science
Foundation coined this term called cyberinfrastructure,
and the idea was that it was supposed to be more
generally relevant to everybody. It was supposed to, you know,
a tie that lifted all boats, enhanced the capabilities
of all sorts of researchers, and the NSF didn’t actually
define cyberinfrastructure. It was clear that if you
wrote a proposal to the NSF, you had to use the word
cyberinfrastructure, and the NSF always
danced around it. They’d say things like, “Cyberinfrastructure
isn’t just,” “Cyberinfrastructure
is more than,” “Cyberinfrastructure
is not limited to.” There was never a sentence that
came out of the NSF that went, “Cyberinfrastructure is.” So, Steve Sims and I and several of our friends coined
this definition of cyberinfrastructure, and it has become the
standard definition. And, I won’t read the whole
thing to you, but the gist of it is, “Computing systems,
data resources, people, software, networks,
all tied together to improve research productivity
and to enable things that would otherwise
not be possible.” And, in this new environment
of cyberinfrastructure designed to be broadly useful to all of researchers throughout
the U.S., we really thrived. 2003, we received funding
to be the informatics core for a collaborative initiative
studying fetal alcohol spectrum disorder. In 2004, IU got another $53
million from the Lilly Endowment for something called
METACyt which was designed to improve biological research
on the Bloomington campus. We received four and a
half million more dollars from the NSF to be
part of the TeraGrid. We received $1.7
million from the NSF to create something
called the data capacitor. This was an episode
in sleep deprivation. There was, like, three months
in which we wrote just around, just around $60 million
worth of grant proposals. Every last one of
them was funded. Boy, we were just
stupid tired by the end, but we got some funding for
some really great projects. The data capacitor in
particular stands out, and Steve Sims was really
the architect of this. We decided there were better
ways to move data around than by strapping your
data to burrows. The data capacitor was designed
to be very large and very fast for collecting data and then
dispersing it where you wanted to put it to have
it be analyzed. Oh, yeah, in 2005, on April
Fools’ Day, I was promoted to the Associate Dean for
Research Technologies. And, in 2006, we unveiled
the big red supercomputer which debuted in 23rd
place on the top 500 list. This is a list of the 500
fastest supercomputers in the world. It’s the highest we’ve
ever been on that list. It was just an amazing
feat of computer design and implementation, and boy
that supercomputer was fast. We also, in Indianapolis
installed a BARCO virtual reality theater, and immersive,
3D stereoscopic environment like the CAVE, only better. 2007, we helped Geoffrey
Fox get nearly $2 million to study arctic and
Antarctic ice sheets. 2008, we received the
second round of funding to convert the pervasive
technology labs into the Pervasive
Technology Institute. 2009, with a considerable
amount of help from us, Geoffrey Fox received
just over $10 million to create an experimental grid and computer system
called FutureGrid. 2010, we partnered with a
private company called Whamcloud to improve the software that runs the data
capacitor then and now. 2011, notably, we received
funding for the National Center for Genome Analysis Support. This was another big step for
us in getting federal funding to support federal, national
research communities. So, the National Center for Genome Analysis Support
supports any researcher in [inaudible] doing classified
research who needs help with genome assembly or
transcript dome assembly. And, in 2011, we were
funded to be a major partner in something called XSEDE. It’s a terrible name. I know. Everybody knows. It’s been a terrible name
since its conception. It stands for the
Extreme Science and Engineering Discovery
Environment. It is the coordinating and
support function funded by the National Science
Foundation to support all of its supercomputers
and cloud computers. And then, in 2012,
we got $3.2 million to operate something called
the Open Science Grid. So, the Open Science Grid is
the very large grid of a lot of computers, particularly
designed to analyze data from Large Hadron
Collider at CERN. And, 7 by 24 by 365 that
system was kept up and running by people in Bloomington
and Indianapolis. So, at this point, Indiana
University is a major player on the national scene
in research computing. We’re involved in XSEDE. We’re involved in the
Open Science Grid. We’re getting lots
of federal money. We’re doing lots of very, very
cool stuff, but we weren’t done. In 2013, we dedicated
Big Red II, the first one PetaFLOPS
supercomputer owned by an individual university. That’s 1000 trillion
mathematical operations per second. Also, the only big, only
supercomputer dedication that I can think of
that was attended by an Oscar winner
and a Grammy winner. And, that computer
is with us today. It is the backbone of our supercomputing
environment here at IU. In 2014, we received a grant from the National Science
Foundation to create Jetstream which was the first
NSF funded cloud system for general purpose research. So, here, we’ve gone
beyond the first thing that some university can
afford to beating the major, national supercomputing
centers to get NSF funding to run a cloud resource for the
National Research Community, and we named it by analogy
with the, with the Jetstream in the upper atmosphere
which is at the boundary of two larges bodies of air
up in the upper atmosphere. The idea behind Jetstream
is that it was to sit at the boundary between the
existing NSF funded computing infrastructure and
its users and the rest of the research community that
was not using NSF funded systems because the needs of these
people were not well met by the existing NSF
supercomputers. And, we succeeded. Something over 80% of the
people who get accounts on Jetstream are getting
their first account on an NSF funded
computing resource. We have more biologists
using Jetstream than any of the other NSF
funded computers. And, major, major use in other
areas of science, geology and atmospheric science
in particular. Yeah, and we just sort of,
the hits keep, kept on coming. So, that’s been kind
of a quick whirlwind, here are the grants we got. Here are the things we did. I want to talk a
little bit about some of the persisting impact we’ve
had in research scholarship and creative activity. We’ve supported a
bunch of cool research. There was an article just a few
years ago by David Polly of IU and one of his colleagues at
University of Nebraska Lincoln that proved that
everything we thought we knew about the evolution of limblessness in
snakes was wrong. That research was done
using our Corey computer. We’ve made important
contributions to the understanding of fetal
alcohol spectrum disorder. We had supported data
analysis from one of the best earth based
telescopes in existence. This M51 galaxy picture here was
taken with the one degree imager out at Kitt Peak in Arizona. Just a lot of really
cool science. A lot of impact on the arts. Amongst fine artists, one of
the first people in the U.S. who had virtual reality
as their primary means of artistic expression
received tenure at IU, Margaret Dolinsky who’s one of
our collaborators and partners and is shown here on the left. Nicole Jacquard sculptor. Uses 3D printing in her art. Rob Shakespeare has been
one of the big users of our supercomputers
from the very inception, and here is a picture of the
light totem which is really one of the top items if you
tour IU Bloomington. So, we’ve supported some
very, very cool artistic work. We have really, really
impacted the university’s access to visualization systems. So, this right here is
a prettied up John-E-Box that is photographed in
the Indianapolis Museum of Art showing some of
Margaret Dolinsky’s art. This is a now retired
visualization system that was up in Indianapolis that
allowed you to look at very, very large, very
detailed images. This picture here is of
something called an IQ wall. It’s part of the series
of technology developed by the advanced virtualization
lab. There’s an IQ kilt
out in the atrium. This is being used in a
lecture about a collection of photographs taken in
South Africa in the 1950. This picture doesn’t
do it justice. When you see the
photographs of the conditions under which minors worked in
South Africa in the ’50s blown up to six fee by eight feet. It’s, there’s just no
good words to describe it. And, that is made
possible, that sort of impact of that photographic art is made
possible by the Immersadesk. This is an immersive
visualization system that’s right next door in
the Innovation Center. This is me and Governor Daniels in the BARCO VR theater
up in Indianapolis. We are looking at the
results of the simulation of proton neutron interactions on the surface of
a neutron star. So, really cool stuff in the
visualization environment. Viz deserves more
than one slide. This thing here is called
the science on a sphere. It’s right outside
this lecture room. It’s great for any sort
of data that you have that you can project on
a giant ping pong ball, pictures of the surface of
the Earth, ocean temperatures, the surface of the Earth,
the world’s entire history of the balls used in World
Cup soccer championships. Cool stuff there. Also, Chauncey Frend
has invented a very, very cool immersive VR
environment called PIPES. You sit in it with VR
goggles, and you get scent and temperature and air flow. So, when you walk outside
into a virtual reality garden, you smell the flowers, you feel
the sun, and you feel the wind. It is a very, very
intense immersive virtualization environment. That’s Chauncey right there, and he does have a startup
that’s commercializing this technology. Performing arts. Werner Herzog was
here a few years ago. Research technologies, advanced
visualization lab has been very, very involved in the IU cinema. When Herzog was here, AVL created the digital
cinema packages for all of the Herzog movies
that were screened. That’s essentially a
digital compilation of everything you need
to both show the movie and understand it. And, those become persistent
artifacts you’ve done that, and that’s what everybody
else uses. We have enabled an
internet required opera about global warming. Now, you may not have even
known that there was an opera about global warming,
but there is. It’s called Auksalaq. It was written by Scott
Deal, percussionist of IUPUI and one of his collaborators. And, the only way you can
experience it is on the internet because it is explicitly
designed to be performed by multiple groups of performers
in multiple locations. And, the only way to
experience the whole thing is to throw a headset on
and watch it one the web. So, some really cool arts. Digital humanities. Research technologies has
gotten very, very involved in what’s called photogrammetry
where you take lots of 2D pictures of 3D objects and then you create
a 3D reconstruction. So, here we have the mermaid
from Showalter Fountain. We have here a 3D image of a
pre-Columbian settlement and set of religious buildings in Mexico
that was done with thousands of 2D photographs that were
taken from a little drone and then stitched together. So, you can fly through this
in virtual reality in 3D or you can print it out with
a digital printer and see it as a big piece of
starch as well. Digital humanities. I’m going to come back
and talk a little bit more about this topic a couple
minutes, but this is an analysis of word use in Kurt
Vonnegut’s novel “Cat’s Cradle”. I’m not a Vonnegut expert. I’m certainly not an
expert on word analysis, but what this graph shows
is the relative use of a set of key terms over the
course of this novel that helps writing
scholars and scholars of English understand
what Vonnegut was doing with this book. And, I’m going to ask the
very quick question how do you suppose it was that
people got a hold of a copy of a copywritten book
that they could, then, analyze with computers? I will answer that question
for you in a few minutes. So, I’m going to do
a quick rewind here, and I want to talk
just a little bit about the Pervasive Technology
Institute in isolation of research technologies and the
other things I’ve talked about. So, PTI has sort of grown up. It’s persisted from
1999 to today. Has a group of centers
that have done some very, very important research,
and like I said, importantly, have persisted. And, very, very quick
history, established in 1999, second round of funding in 2008. And, in 2008 even, it was still
sort of very inwardly focused. We got money from
the Lilly Endowment. We’re going to go
spend that money. We’re going to do good research. And then, in 2017, there
was a really significant organizational change. My job got split
into two pieces. My old job learning research
technologies, at that point went to Matt Link which
I’m very happy about. And, I’m focused exclusively on the Pervasive Technology
Institute, and I just want to run through the centers and say a little bit
about what they do. The Center for Applied
Cybersecurity Research, the, one of the oldest of
the PTI centers and one of the most important. We live in a dangerous world. The world is filled with
bad actors who want you to believe fake things
on Facebook. They want to mess with
our ability to do research by messing with people’s
research data. May want to hijack your
accounts at your bank. CACR works with funding from a
variety of government sources to help secure critical
military infrastructure, scientific infrastructure, and our everyday
computing infrastructure. So, if you’ve ever listened to a
moment of security on the radio, that is a CACR product,
and they serve the nation in improving the security of our
networked computing resources with a very strong focus on
research computing resources. The Science Gateways
Research Center. The Science Gateways Research
Center creates technology that isn’t at all obvious
to see, but what it does is that it ties together
multiple supercomputers, multiple data storage systems,
visualization software, and it makes running a very, very complex engineering
analysis, data analysis, or computer simulation seem
very simple through a web portal because they’re very, very
sophisticated web portals. So, people who are
every day, ordinary, practicing scientists can
do chemical engineering and docking potential
drugs against drug targets. We’ve got some very, very sophisticated
educational gateways, and these things are created by the Science Gateway
Research Center. The National Center for Genome
Analysis Support has helped researchers all over the
U.S. assemble the genomes and transcriptomes of a bunch
of organisms ranging from things that are economically important. A lot of these, you know, cows are economically
important diatoms, ocean invertebrates very
important, very cool. You know, very cool
tiger salamanders. This is just a sampling of
the organisms whose genomes and transcriptomes have
been assembled with help from the National Center
for Genome Analysis Support. The Data 2 Insight Center lead
by Beth Plale who’s currently on leave with the National
Science Foundation is focused generally on turning
data into insight. One of the things that’s
happening right now is that the Hathi Trust Research
Center is being carved out from the Data
2 Insight Center and is being created
as its own center. The Hathi Trust Research Center
is related to the Hathi Trust, the place that has copies
of all of the Google Books from the Google Books project. Sixteen million volumes of
text in a variety of languages, ten million of those 16
million volumes are protected by copyright. So, I asked the question,
you know, how do you analyze one
book that’s in copyright. You know, you can analyze
one by buying a digital copy and having it scanned. And maybe you can
do that with ten. But, if you want to analyze
word usage and look at, say, 1000 or a million or five
million books that are protected under copyright and you
don’t want to be sued, it’s a very hard thing to do. What the Hathi Trust Research
Center creates is something called a data capsule
that is modeled after Social Science
Research Centers where people analyze
very, very sensitive data. So, all that copywritten text is
kept inside that data capsule, and when you want to analyze
it, you send your query into the data capsule. And, the Hathi Trust Research
Center computing tools do the analysis, and they send the
results back out without sending out enough text to
violate copyright rules. And, this system is secure
enough that it is approved by all of the lawyers at all
of the big publishing companies that have texts that are in it. So, it makes possible
humanities research that would be impossible without
a way to electronically analyze and manage copywritten text. The digital science center
is led by Geoffrey Fox, School of Informatics,
Computing, and Engineering. They do work primarily
in the area of developing software
for big data. And, collectively, RT and
PTI have aided the economy of Indiana, and the
research scholarship in creative activity
economics of IU. And, I want to talk a
little bit about money for a couple minutes here. We’ve been running this positive
feedback loop where we go out and we get grant awards. We do good stuff. We hire smart staff. Then, we go write
more grant proposals. We get more money,
hire more smart people, do more good stuff. We’ve been running this positive
feedback loop for decades. It has worked really well. Just as one indication of how important our computing
resources are to the community, there are 134 different
departments at Indiana University
where some research in that department
makes some use of our computational systems. So, this pretty diagram
here shows usage of Big Red II in calendar 2017. The size of the box
corresponds to the amount of computing use a given
department has made. So, the biggest box is physics. And then, there’s chemistry,
but there’s biomedical research. Here’s geology. There’s medical and molecular
genetics in Indianapolis. There’s interior design and apparel merchandising,
history of science. All sorts of departments make
use of our supercomputers, 134 departments,
tremendous reach within the university community. Much more than at
our peer schools and peer computing
organizations. We have a tremendous impact
on the grant competitiveness of IU faculty members. So, in 2017, IU researchers
brought in a total of $508 million of external
grants and contracts. If you ask, okay, of that $508
million, the people that brought in that money, do they use our
cyberinfrastructure systems or not? And, the answer is that people
that use our systems brought in $409 million of the $508
million total grant money brought into Indiana University. And then, if you ask
the money, by the way, if you ask the money question. Is it cheaper to
have our own stuff or to buy it off the market? We did a cost comparison
where we pretended that Amazon web services
was as good for supporting computing
research as Big Red II is. By the way, it’s not. For technical reasons that I
won’t go into, having to do with internal networks and this. But, if you pretend
that Amazon is as good for supporting IU
research as Big Red II is, our own systems are still two
to three times less expensive than buying resources
off of Amazon. So, we’re helping IU
competitively win federal grant monies, and we’re doing
it cost effectively. We’ve created hundreds of
years of jobs in the central and south central Indiana area. So, this graph here
shows growth of IT staff in research technologies
and PTI over time. This is base funded staff and research technologies
right here. This is grant funded staff
in research technologies. This is grant funded
staff in PTI. So, dozens of people, hundreds
of persons years of employment at Indiana University in
Indianapolis and Bloomington, and a lot of these people served
the national research community and were very careful
that between eight and five we serve the national
research community equitably. But, the people serving that national research community
live in this community. So, they run into
researchers and students in the grocery store,
in meetings, at lunch, at the Binford Rogers
Spring Festival. I once did two hours
of consulting at my granddaughter’s spring
festival at Binford Rogers because I ran into a
professor, ran into a professor at the spring festival
who was there with his kid who was having a problem. And, while his kid and my granddaughter
were off winning cakes at the Binford Rogers Festival, we were spending two hours
solving one of his problems. So, the fact that these people
supporting the national research community are here makes a
difference to the quality of the IT resources and skills
that IU researchers, staff, and students have
accessible to them. We have done a lot to
inspire the next generation of technologists. This is a picture of robot camp. Hundreds of students have
gone through robot camp over, well over a dozen years. They come in here. This is a picture
taken in this room. They come in here, they
take, spend a number of days learning how to program
robots, and they go out inspired to pursue careers in technology. So, it’s a really
great thing we’ve done for the future economy
of Indiana. Along the way, we have
supported research that has led to three Nobel Prizes, two in
2013 through our involvement in the TeraGrid and XSEDE. We supported the 2013
Nobel Prize in Chemistry through our involvement
in the Open Science Grid. We supported the 2013 Nobel
Prize in Physics which was given out for the demonstration
of the Higgs Boson, also through the
Open Science Grid. We supported research in
verifying the existence of gravitational waves that led to the 2017 Nobel
Prize in Physics. This, by the way, is a picture from the Large Hadron
Collider at CERN. And, it’s really big. To give a sense of how big,
you know, this is a big, a big wire spool here. These were essentially,
you know, human size guard rails there. And, a lot of this work that
led to Nobel Prizes was research that was done at
the LHC and analyzed with computing systems
that we run. So, quick rewind to talk
about national strategies in advanced computing. In the 1940s, the purpose of
the fastest supercomputers in the U.S. was really clear. They were defense instruments. This is ENIAC. This system was used to calculate artillery
trajectories, and it was also used at one of
the first feasibility studies of thermonuclear weapons. In the 1960s, our national
priorities in science and technology were
really clear. They were set by
President Kennedy. Put a man on the moon, and that
was seen as a symbolic effort in the international campaign
for the hearts and souls of people throughout
the world in the battle between democracy and communism. In the 1980s, there was a lot
of clarity in computing as well. This concept of the time
machine, the Department of Energy supercomputers were
the fastest in the world, and oh, yeah, there
were all sorts of industrial uses as well. Building a big bulldozer
is expensive, and when a big bulldozer
tips over, it’s really bad. And, this bulldozer
is really big. That’s, you know, that’s a
person sized door right there. Caterpillar simulated this
bulldozer at supercomputers of the National Center for
Supercomputing Applications in Champaign-Urbana, Illinois
and simulated it to the point that they were sure
they knew how to build it before they ever
built their first prototype. So, lots of great things
in the ’80s, and also, during the ’80s something,
you know, something was just
generally true was that new processors were
built for supercomputers first and then filtered out
into the consumer market. In the 1990s, we had the
Comprehensive Test Ban Treaty, and supercomputing was viewed
as, again, a defense asset. The idea was that we would
out compute the Russians, and we would be able to maintain
our nuclear deterrents more effectively than the Russians
in the absence of testing because we could
simulate nuclear weapons and even build nuclear weapons
without ever once testing them as a result of our
superiority in supercomputing. That sort of began to
change in the 2000s when, for the first time,
somebody other than the U.S. had the fastest
supercomputer in the world. The Japanese Earth Simulator
was actually the number one supercomputer in the world for
about two and a half years. And, there’s about one
week of shock in Washington in which people said,
“Oh, my god. We have lost leadership
in supercomputing. This is really bad. We need to do something
about this.” And, after about a
week, people realized that it was actually
cheaper to say, “You know, the Japanese supercomputer,
it’s really special purpose. It isn’t generally useful. We shouldn’t worry about it. The fact that it’s number one
on top 500 list is an anomaly. We’re still number one.” There’s even less clarity now. We had this National Strategic
Computing Initiative signed by President Obama in
2015 which was designed to put the U.S. back
in a position of global leadership
in supercomputing. But, in fact, we’re not. We’re distinctly
second, behind China. China has and has had the
fastest supercomputer on Earth, at least that we know about,
in terms of the top 500 list. If you look at the top
ten fastest supercomputers on that list, China has a larger
fraction of the total capacity in the top ten than
the U.S. has. If you look at the capacity
of the top 500 list, China has larger capacity total
of its supercomputers combined than the U.S. has, and China
has more supercomputers on the top 500 list
than the U.S. has. And, China is going to beat
the U.S. to an exaflops. That’s 1000 petaflops or 1000
thousand trillion mathematical operations per second. China’s going to beat the U.S. to having an exaflop
supercomputer probably by two years. The DOE announcements
recently notwithstanding. Unless there’s something that
wasn’t declared publicly as part of the DOE strategies, the DOE
strategies that are announced at least will, in fact, not
put the U.S. back in a position of national, of international
leadership in supercomputing. It will just keep U.S.
number two behind China. So, we have really
lost what we thought of as being a key
U.S. strategic asset that we held onto for decades. And, a lot of other
things changed. In processor development,
it’s now the commodity market that is dog and supercomputers
that are tail. So, rather than having
processors built for supercomputers and bleed
out into the consumer market, we have processes that are
built for the consumer market and then adapted
to supercomputing. The longstanding trend of
processors getting smaller, faster, and cheaper
is coming to an end. After years of pressure
from the NSF in particular for computer facilities and
computer intensive research to be more generally useful. There’s a lot of pressure now to very quickly say what
you’re going to do and how long and what the practical
benefit is going to be. And, there’s less
federal funding. And, it’s worse than that. State funding for universities
and colleges is going down. In the next few years
hundreds of institutions of higher education are
going to go bankrupt, smaller colleges
and universities. We’re facing worse and very
different security threats than we ever have before. The one thing that hasn’t
changed is there’s still no money from the federal
government for arts and humanities. And so, we’ve got some real
challenges moving forward at the university level
and at the national level. In terms of IU funded resources,
one of the things we have to do is just continually
demonstrate the value of what we do to the university. We have to demonstrate
the financial soundness of our strategies, and
we are working very, very hard to broaden the scope
of utility of the facilities that we provide to
the university. In terms of federal
funding, we’ve just got to work hard and adapt. There’s less federal money,
and getting grants is harder. And, we just have to
work really, really hard. And, we’ve got work in
society to do as well because there are very
difficult things happening at the national level, and
they’re not going to get better without smart people
being engaged and involved in those conversations. I want to talk a little bit
about some of the things that we’re doing
specifically along the lines of what I just described. Research technologies,
Matt Link, Robert Henschel, Dave Hancock, and their
colleagues are developing this thing called RED, the Research
Desktop, which is designed to make it even easier
than it already is to use IU supercomputers. So, you get something that looks like a standard practical
user interface you might see in a laptop or a
desktop computer. Only, it’s running
IU supercomputers and massive data
storage systems. We are at the other end
supporting IU’s grand challenges with the new supercomputer
called Big Red II Plus which is dedicated to IU’s grand
challenges in precision health, preparing for global change, and
battling the addictions crisis. As well as being used to
address U.S. grand challenges in exascale computing,
potentially working with faculty members in
the School of Informatics, Computing, and Engineering
to build software that will enable the U.S. to be
back in a leadership position in exascale computing. Federal research funding exists to help the nation
support doing good things. In 1945, when there was a
debate going about the creation of the National Science
Foundation, James Bryant Conant, who later, who was at
that point the President of Harvard University,
wrote a letter to the editor of the New York Times. It was really a promotional
piece lobbying for the creation of a National Science
Foundation. And, I have taken the liberty
of updating his language. He wrote, “There’s
only one proven method of assisting the
advancement of pure science, that of picking people of
genius, backing them heavily, and leaving them to
direct themselves.” The Research Technologies
Division of UITS backs the people of
genius in the IU community, backs them strongly,
and trusts them to be doing great
and important work. Conant also said, “There’s
only one proved method of getting results
in applied science, and that is to take people of
genius, back them strongly, and keep them on target.” That’s really what the Pervasive
Technology Institute does. We are a group of
centers that do things that the federal
government wants done, and we do them the way the
federal government wants them done. And, Conant didn’t say a
darn thing about humanities and the arts, but
Research Technologies and PTI both support
humanities and the arts as an important part of
what this university does. And, I already said that. So, the stuff on
this slide I said. I want to talk a little bit about Indiana University
in grant getting. Indiana University bring
in a fair amount of grants. It underperforms on large grants as compared to many
of its peers. One $10 million grant
award is worth more than ten $1 million grant
awards in terms of its impact because at $10 million, you’ve got multiple faculty
members working together. You’re working on a really
big and important problem. You’re doing a lot
of dissemination. You’re telling a lot of people
about what you’re doing. And, you know, IU’s
performance in the ten million and up category is an issue. In the past 15 years,
there have been a dozen of these national core awards
that tend to be very large to the School of
Medicine in Indianapolis where they’re doing some
sort of service on behalf of the national research
community. We’ve got one in
Bloomington, the National Center for Genome Analysis Support. We don’t have any of the
large NSF center grants, the so-called science and
technology center grants. If you look in the last 15 years
for awards that were greater than $10 million, there’s 21 of
them to the School of Medicine. There’s two to PTI. There’s one other to OVPIT, and
three others to Bloomington. There are a few more large
ones if you add up sequences like our involvement
in TeraGrid and OSG. But, when it comes
right down to it, the place that Indiana
University underperforms in getting large grants
is IU Bloomington. And, I think one of
the real challenges is that responsibility center
management creates barriers to creating the very,
very large collaborations that make these large
grants possible. Where we are right now is the
grand challenges program is IU’s best and most important
attempt to facilitate real, large scale collaborations, and
PTI and RT are working very, very hard to support
the grand challenges. And, PTI, in particular,
is just doubling down on getting grant
funding to do things that are both important
at the national level and consistent with
IU’s mission. So, we go after things that
are consistent with IU values and important to the
federal government. And, we’re also working
right now to look for nonfederal funding sources. We have historically
done really well at keeping staff
members employed, keeping people going year
after year after year after year on grant money. And, and we’ve slipped. I used this picture for
many, many years now. This is a picture from a dude
ranch in upstate New York with one guy doing what’s
called Roman riding. He’s got one foot on each of
two horses, and I like to think about our funding support
as having, you know, being comprised of two
horses, the horse of IU funding and of course, the
federal funding. And, the temptation is to focus
on the horse of federal funding because that’s the one that
feels like it’s going to fall out from underneath you. But, what we believe is if we
lost all of our federal funding and all we had was
our IU funding, it would hurt like the dickens. Lots of people would
lost their job, important stuff would go undone,
but you would still be able to recognize the
organization that we are. On the other hand, if we lost
the support of the university, what would be left would
not be recognizable and it would not
be a service to IU. So, we try to pay really close
attention to both horses, and, for the first time ever, the horse of federal
funding has stumbled. We have had a group that lost
its funding, and we lost it, we lost funding for this
group of about five people not because we wrote a bad proposal,
not because we did a bad job. We lost funding for
this group of people because the government shutdowns
that have plagued Washington in the last year
made it impossible for the federal funding
agencies to put out the grant solicitations
to which we planned to respond and keep this group funded. So, through no overaction,
the simple interruption of government functions
has wiped out a group of research support
professionals at Indiana University
who will find other jobs. But, once that group is gone, there’s no putting it
back together again. And, this is just one example of
how precarious we are in terms of our national dialog
and scientific research in the U.S. And, that national
dialog isn’t going to get better without our participation. You know, it’s clear
to me that many of the ideals behind
the computer revolution, the internet, the worldwide
web, a lot of the ideals under which these things were
created have now been subverted. “On the internet,
nobody knows you a dog,” used to be really funny. On the internet, nobody knows
that you are a professional, Russian troll is a very,
very dangerous thing. And, I saw a poster that said, “Misled people make
misled decisions that benefit the
misleaders not the misled.” And, divisiveness pays. I would argue that
Breitbart News and HuffPost have essentially
identical business models. What Breitbart and HuffPost both
do is they scream to group A about group B. The only
difference between HuffPost and Breitbart is
which group is group A and which group is group B. And,
as people who think and believe in the value of truth, these
discussions that are going on in the national
level aren’t going to get better unless we’re
willing to participate in them, and I think as computer people,
there’s a particular obligation on us as well to think about what the computer
revolution has done for the economy of the
U.S. There are countries, and Germany is one of them,
that have chosen high tech and high quality employment. That doesn’t seem to be the
choice that the U.S. has made. Now, what’s our role in that? We need to be, we
need to be part of these national discussions
because they aren’t going to go well without us. I want to say just a very
few things that I learned from Chris Peebles about
living in challenging times. Trust in the excellence of the
faculty of Indiana University. They really are excellent. Some of them are a bit goofy. Some of them are
a bit difficult. On average, they’re excellent. By serving them, we serve some of the best intellects
in the world. Be human, and to be
human is to be kind. And, don’t take it
all very seriously. Chris used to say, “I am
sometimes thrown to the faculty as one might throw
poisoned meat to wolves. They chew on me awhile, eventually find me
distasteful, and throw me back.” And, Chris was a great leader
who believed in the value of the university and didn’t
take it too seriously. I want to say one comment
about my path forward and that research technologies. So, last fall, I gave up
two pieces of my old job. Matt Link became the
Associate Vice President for Research Technologies. Dave Hancock became the
Principle Investigator of Jetstream, and I
am not focused on PTI. It is with utmost confidence
that I look at Dave and Matt as leaders of Jetstream and Research Technologies,
generally. And, I’ve got my work cut out
for me for the next few years. There are tremendous
challenges ahead of us. One of the challenges
that is given to us by President McRobbie is to help
Indiana University become one of the great universities of
this century, and I believe that we are up to this task. And, the things that we do in
research technologies, PTI, and university information
technology services generally will help IU become one
of the great universities. Of this century. And, there you have it. We have changed the
world some already. We just need to keep doing
it, and we need to do it more. With that, I want to
thank a bunch of people. The Lilly Endowment, boy, if you
want to look at one organization that got us from where we were
in 1996 to where we are today, the repeated injections
of thoughtful support from the Lilly Endowment have
been absolutely critical. We’ve done a lot of research
with a lot of support from a bunch of federal
funding agencies. All opinions presented here
are those of the speaker and may not necessarily
reflect the opinions of any federal funding agencies, and certainly some things I
have said would be disavowed by essentially any federal
funding agency, I am sure. And, I want to thank
a bunch of people. I want to thank my
family, my wife Marion. Kai, George, Madeline, Tony,
Kristan, and Elan, and Noah. Where’s Noah. Yeah. So, thank you
all very much. I want to thank all
of my colleagues. One of the things that
I have always said is that I never mistake
the leader for the team. When I was the leader of
research technologies, I was just the leader. It’s the team that matters. And, I want to thank
everybody in our team, PTI past and present, but especially
the sort of the brain trust. Eric Wernert, Therese
Miller, Steve Sims, Matt Link, Dave Hancock, Robert Henschel,
Von Welch, Marlon Pierce, Mike Boyles who did a lot of the
work behind a lot of the systems that we’ve talked about. Early leaders Mary
Papakhian and Melinda Husk. I want to thank Winona
Snapp-Childs who edits me better than any other editor ever has. The people in OVPIT who
supported me particularly in the last 14 months as I
have been overcoming cancer. The largest care package I got
was from the Finance Office. [ Laughter ]>>Telling.>>You know. [ Laughter ] The Finance Office. Now, I will say that they got,
they got advice from Matt Alan. So, they got good advice,
but Finance Office. Richard Meraz has been a great
help to me in this time as well. Rudeana Honeycutt,
where’s Rudeana. Rudeana, stand up. [ Applause ] Rudeana has ruled my
days from eight to five and when I have traveled
for just about 13 years. We have come to the conclusion
that no human being deserves to have to be Matt
Link’s boss and my boss. So, Matt gets to keep
Rudeana, and I don’t, and I thank Rudeana many years. I want to thank the faculty,
staff, and students of IU. A services organization
is worthless without people to serve. The IU community gives
us great people to serve. And, I want to end by
thanking my mentors, particularly Chris Peebles, Brad
Wheeler, and Michael McRobbie. And, with that, are
there any questions? Thank you all very much. [ Applause ]>>I loved the talk
because, okay. Thank you. I loved the talk because I, it
reminded me of when I joined up which was around
the AVIDD time. Remember this? And, I don’t know if it came
through for everybody else, but it sure came through for
me as it was a lot of fun, man.>>Oh.>>We had, we had some fun. So.>>We had, we have done
some awesome and fun work, and we’ve invented
a bunch of stuff. It’s been really cool, and one
thing I should say is Scott was involved in building one piece
of the Large Hadron Collider. So, the senior most
person involved in the Large Hadron Collider within the U.S. was Harold
Odring [assumed spelling] Physics professor at Indiana
University, and one part of one of the two experiments that
constitutes the major reason for the existence of the Large
Hadron Collider was constructed on Third Street in Swain Hall. So, our involvement in
the Open Science Grid and the Large Hadron Collider from the computing standpoint
very definitely drafts the leadership and excellence of the
physics community here at IU. And, it has been a lot of fun,
and it will continue to be fun because I am not done yet. So. Other questions? With that, let me thank
you all for your attention. This will be the last time ever
that I give a talk like this because from now on,
talking about the history of RT is Matt’s job
and other people’s in the Research Technologies
job. So, thank you all very much for your attention,
and have a great day. [ Applause ]