[House Hearing, 113 Congress]
[From the U.S. Government Publishing Office]
DATA CENTERS AND THE CLOUD, PART II: THE FEDERAL GOVERNMENT'S TAKE ON
OPTIMIZING NEW INFORMATION TECHNOLOGIES OPPORTUNITIES TO SAVE TAXPAYERS
MONEY
=======================================================================
HEARING
before the
SUBCOMMITTEE ON GOVERNMENT OPERATIONS
of the
COMMITTEE ON OVERSIGHT
AND GOVERNMENT REFORM
HOUSE OF REPRESENTATIVES
ONE HUNDRED THIRTEENTH CONGRESS
FIRST SESSION
__________
JULY 25, 2013
__________
Serial No. 113-52
__________
Printed for the use of the Committee on Oversight and Government Reform
Available via the World Wide Web: http://www.fdsys.gov
http://www.house.gov/reform
----------
U.S. GOVERNMENT PRINTING OFFICE
82-434 PDF WASHINGTON : 2013
-----------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office Internet: bookstore.gpo.gov Phone: toll free (800) 512-1800;
DC area (202) 512-1800 Fax: (202) 512-214 Mail: Stop IDCC,
Washington, DC 20402-0001
COMMITTEE ON OVERSIGHT AND GOVERNMENT REFORM
DARRELL E. ISSA, California, Chairman
JOHN L. MICA, Florida ELIJAH E. CUMMINGS, Maryland,
MICHAEL R. TURNER, Ohio Ranking Minority Member
JOHN J. DUNCAN, JR., Tennessee CAROLYN B. MALONEY, New York
PATRICK T. McHENRY, North Carolina ELEANOR HOLMES NORTON, District of
JIM JORDAN, Ohio Columbia
JASON CHAFFETZ, Utah JOHN F. TIERNEY, Massachusetts
TIM WALBERG, Michigan WM. LACY CLAY, Missouri
JAMES LANKFORD, Oklahoma STEPHEN F. LYNCH, Massachusetts
JUSTIN AMASH, Michigan JIM COOPER, Tennessee
PAUL A. GOSAR, Arizona GERALD E. CONNOLLY, Virginia
PATRICK MEEHAN, Pennsylvania JACKIE SPEIER, California
SCOTT DesJARLAIS, Tennessee MATTHEW A. CARTWRIGHT,
TREY GOWDY, South Carolina Pennsylvania
BLAKE FARENTHOLD, Texas MARK POCAN, Wisconsin
DOC HASTINGS, Washington TAMMY DUCKWORTH, Illinois
CYNTHIA M. LUMMIS, Wyoming ROBIN L. KELLY, Illinois
ROB WOODALL, Georgia DANNY K. DAVIS, Illinois
THOMAS MASSIE, Kentucky PETER WELCH, Vermont
DOUG COLLINS, Georgia TONY CARDENAS, California
MARK MEADOWS, North Carolina STEVEN A. HORSFORD, Nevada
KERRY L. BENTIVOLIO, Michigan MICHELLE LUJAN GRISHAM, New Mexico
RON DeSANTIS, Florida
Lawrence J. Brady, Staff Director
John D. Cuaderes, Deputy Staff Director
Stephen Castor, General Counsel
Linda A. Good, Chief Clerk
David Rapallo, Minority Staff Director
Subcommittee on Government Operations
JOHN L. MICA, Florida, Chairman
TIM WALBERG, Michigan GERALD E. CONNOLLY, Virginia
MICHAEL R. TURNER, Ohio Ranking Minority Member
JUSTIN AMASH, Michigan JIM COOPER, Tennessee
THOMAS MASSIE, Kentucky MARK POCAN, Wisconsin
MARK MEADOWS, North Carolina
C O N T E N T S
----------
Page
Hearing held on July 25, 2013.................................... 1
WITNESSES
Mr. David Powner, Director, Information Technology Management
Issues, U.S. Government Accountability Office
Oral Statement............................................... 6
Written Statement............................................ 9
The Hon. Steven Vanroekel, Acting Deputy Director for Management,
Federal Chief Information Officer, Administrator for E-
Government and Information Technology
Oral Statement............................................... 58
Written Statement............................................ 60
Mr. David L. McClure, Associate Administrator, Office of Citizen
Services and Innovative Technologies, U.S. General Services
Administration
Oral Statement............................................... 64
Written Statement............................................ 66
APPENDIX
Questions for the Record......................................... 88
DATA CENTERS AND THE CLOUD, PART II: THE FEDERAL GOVERNMENT'S TAKE ON
OPTIMIZING NEW INFORMATION TECHNOLOGIES OPPORTUNITIES TO SAVE TAXPAYERS
MONEY
----------
Thursday, July 25, 2013,
House of Representatives,
Subcommittee on Government Operations,
Committee on Oversight and Government Reform,
Washington, D.C.
The subcommittee met, pursuant to call, at 9:34 a.m., in
Room 2154, Rayburn House Office Building, Hon. John Mica
[chairman of the subcommittee] presiding.
Present: Representatives Mica, Meadows, Connolly, and
Pocan.
Staff Present: Alexia Ardolina, Majority Assistant Clerk;
Richard A. Beutel, Majority Senior Counsel; Caitlin Carroll,
Majority Deputy Press Secretary; John Cuaderes, Majority Deputy
Staff Director; Linda Good, Majority Chief Clerk; Tyler Grimm,
Majority Professional Staff Member; Mark D. Marin, Majority
Director of Oversight; Sarah Vance, Majority Assistant Clerk;
Jaron Bourke, Minority Director of Administration; Adam
Koshkin, Minority Research Assistant; Safiya Simmons, Minority
Press Secretary; and Cecelia Thomas, Minority Counsel.
Mr. Mica. Good morning. I would like to call this hearing
of the Subcommittee on Government Operations to order.
Welcome, everyone, this morning. The topic of today's
hearing is Data Centers in the Cloud, Part II: The Federal
Government's Take on Optimizing New Information Technologies
and Opportunities to Save Taxpayers Money.
Mr. Issa usually gives a little statement of our mission,
but that title almost sums it up. We are looking and have the
responsibility to review various operations of the Federal
Government and representing the taxpayers; looking for the most
efficient, economical, and responsible means of carrying out
the positions and conducting proper oversight of the agencies
that perform those responsibilities.
Today we have one panel and three witnesses. We welcome
them. And our order of business will be as follows: we will
start with opening statements by members and then we will
proceed to hear from our three witnesses, and after that we
will move to questions. We will hold the questions until we
have heard from all of the panelists.
With that, I will recognize myself for an opening statement
and then turn to our Democrat leader, the ranking member, Mr.
Connolly, for his comments.
I again have to state that our responsibility is that we
look at what the various agencies are doing, in particular,
some of this activity of our subcommittee may not be the flashy
part of serving in Congress, with the hearings and all of the
cameras and all of that, but, nonetheless, this is probably as
important a responsibility as we have. This is the meat and
potatoes, finding out where the money is being spent.
On the particular program of IT, we spend a lot of money.
It is estimated, I think, $84 billion annually. And we have had
witnesses. This is our second hearing. Unfortunately, we didn't
have OMB and GSA representatives at the last hearing; I am glad
they came this time. But because, again, of the sheer size and
scope of this activity, the estimates, again, of potential
savings are maybe as much as 50 percent of what we are
currently saving. If we can consolidate, if we can use the
mechanism of cloud computing, a whole host of efficiencies
brought into this process, we have the potential for saving in
a time when we are approaching a $17 trillion deficit.
Substantial money.
So I again welcome the witnesses. As we face this time of
tight budgets, it has never been more important for the Federal
Government to continue its efforts.
Today's hearing actually looks at, unfortunately, an
attempt by the Administration, dates for several years now in
trying to maximize the return on investment and reduce the
operational risk and provide responsive services to citizens
through some IT consolidations. We, unfortunately, have found
through our investigation that we don't have a pretty good
track record; that maybe the intent was good, but unfortunately
what was set out as some goals and new approaches to achieve
success, have not worked.
We have two charts here that I want to point to, and they
show, unfortunately, failed IT investments since 2003. The long
and the short of it is we have lost about $9.2 billion in those
failed attempts, a pretty significant amount.
The second chart shows the number of troubled IT
investments. That is enumerated on the chart that you see up on
the screen. But this is startling: $102 billion is currently at
risk from, again, a simple evaluation information we have
received. So this is quite troubling.
While GAO had indicated initially that we had some I think
it was 3,133 data centers, the most current data we have
received says that that estimate is some now 7,145. So a pretty
dramatic departure from what was originally estimated as the
number of these non-core centers; and that totals 6,650.
Unfortunately, since 2010 they have closed or consolidated
484 of these non-core centers. At the current rate, and that is
over two years, while they estimate by the end of the year they
will have 855. But they plan, at least the plans we have been
told, to close 3,400, about half of the new estimate non-core.
Unfortunately, if they keep it up at the pace that they have
begun, it looks like it is almost impossible to achieve that
goal. That is part of what we want to hear, how we are
recalculating and how we plan to again reach a significant
goal, which was originally estimated, I think, at some $3
billion. The current calculations indicate that we will be at a
$2.4 billion shortfall.
Unfortunately, the duplications of common IT systems are
pretty widespread, and GAO itself has reported that the
Government funded 622 separate human resource systems at a cost
of $2.4 billion, 580 financial systems at a cost of $2.7
billion, and 777 supply chain systems at $3.3 billion; and the
list continues. Many of these systems perform, unfortunately,
the same function.
OMB has the responsibility to oversee large IT projects,
but often rolling out new and large management initiatives,
when they do with great fanfare, unfortunately it doesn't
appear a lot is done to follow up or implement these programs;
and we are going to look at one example, the Federal Data
Center Consolidation Initiative, which was rolled out again in
2010. As I point out and predicted, they close a lot more of
these centers, 40 percent of them, by 2015 and save that $3
billion. The target is, again, very illusive from what we can
see with the statistics and facts of accomplishment to date.
So OMB, unfortunately, has grossly underestimated the
number of data centers. Now we have to deal with a much larger
number than originally reported. Unfortunately, OMB's savings
to date are minimal, probably about $300 million, or 10 percent
of the promised $3 billion that had been promised.
Unfortunately, we also find that OMB is also skeptical
about their own new process and how it will work. Its own
testimony identifies deficiencies, and we will look at how that
IT Dashboard hasn't worked. We will also have to find out,
firsthand today, how OMB plans to accelerate its use of key
management initiatives and look at how, again, we are going to
achieve these savings. So we want this to be a positive
hearing; find out, again, what has gone wrong, and then see how
we can correct it.
We want to work with OMB and GSA. Unfortunately, GSA, I
found, has failed to develop or roll out its own critically
needed new system to streamline the fashion by which
contracting officers ensure contracts are responsible vendors.
That system, called SAM, System for Award Management, was
completely restructured after multiple failures and millions of
dollars wasted.
And then, finally, what the GSA, I think they set a poor
example themselves with over 109 core data centers. I am told
they have only actually closed one to date. So we will want to
get an update on where one of our principle agencies
responsible for procurement and heavy involvement in the IT
business has failed itself.
So those are some opening commentaries. I have been picking
up from Mr. Issa and also Mr. Connolly their efforts to try to
get this to work better, be more effective, find ways to
consolidate, find ways to streamline and provide the leadership
to make that happen.
So let me yield now to Mr. Connolly, our ranking member.
Mr. Connolly. Thank you, Mr. Chairman, and thank you and
your staff for holding this second hearing on this subject. And
I agree with you, this is the sexiest subject in town. Data
center consolidation, it is amazing the press just isn't here
in overflow, really.
But it actually is one of those building blocks, as you
said, Mr. Chairman, of Government that is actually very
important.
First of all, I want to welcome our witnesses. I am glad
they are here. Some of the things that have happened since our
field hearing, one of which is Mr. VanRoekel is now at OMB. So
that, to me, is a heartening development because I think now we
can better integrate OMB's management responsibilities with the
goals we have here, rather than sort of as a set-aside nice
thing to do. So integrating that into key management decision-
making in the Executive Branch I think is a real step forward.
Mr. VanRoekel, congratulations on that new role. I know
that is going to serve all of us well.
From my point of view, and I certainly am open to testimony
that would suggest otherwise, I share the chairman's concern
that we seem, when we are looking at metrics, and it isn't just
our opinion, it is the Government Accountability Office, we
seem to be falling behind. Good news: we have doubled the
number of companies certified under FedRAMP since our field
hearing, from three to six. But there are a lot of companies
that would like to qualify and that would qualify but for a
rather glacial pace of certification. We want to be thorough,
but golly, gosh, darn, it seems to me we could do better.
There are issues, as the chairman indicated, with the IT
Dashboard. And in terms of data center consolidation, any way
you measure it, we may be more accurate in defining what a data
center is and, thus, the proliferation in numbers that might
seem to suggest we are actually going in the wrong direction
but in fact we are more accurate in trying to capture what is a
data center, and that is good. But if you look at agency
performance, agency by agency, essentially only four agencies
are playing in the game, that is not acceptable. That is not
acceptable.
Finally, I would say to the Administration, and I guess Mr.
VanRoekel, primarily through you as maybe the senior
Administration official here, but this committee, including the
chairman, passed a bill, FATAR. It is the first comprehensive
rewrite of Federal IT acquisition in 20 years. It is designed
to update and, in many ways, sort of replace Clinger-Cohen. It
is the friendliest, most sympathetic bill you are going to get
out of the Congress. It is, in large measure, a codification
of, in fact, initiatives and reforms undertaken by this
Administration, mirabile dictu coming out of this Republican
Congress.
But if the Administration decides to spurn that
legislation, that has passed the House already, I would just
say to you you are going to have problems on both sides of the
aisle. This represents, I think a real bipartisan effort. We
work very hard to try to get this right. We tried to consult
with the Administration. We consulted extensively with
industry.
The chairman, Mr. Issa, could have filed this bill back in
October. To his credit, to the staff's credit, they kept it
open, negotiating with us, with industry, with the
Administration, with others from October until we filed it
finally late in February. And we were willing to take
additional modifications to try to make sure we get it right.
This is a bill designed to try to be helpful, to try to spur
the very reforms undertaken by this Administration, by Mr.
VanRoekel and his predecessor.
So I urge you to go back and consider support for this
legislation. If there are changes, great, but a position of
opposition is not going to sit well in this Congress on both
sides of the aisle. So please consider it the helpful tool it
was intended to be. We think that this is a subject matter that
needs attention. And while it may not draw big crowds, as the
chairman indicated, in terms of potential for savings,
potential for the deployment of technology as a tool, an
instrument for augmenting the decline in resources for so many
agencies is profound; and we need you as a partner and we want
to be a partner.
So, with that, I look forward to the testimony, Mr.
Chairman, and, as you know, I may have to sneak out to manage
an amendment on the floor, but I will be back. And again I
thank the chair for holding this hearing.
Mr. Mica. Thank you, Mr. Connolly.
Other members? Mr. Meadows?
Mr. Meadows. Thank you, Mr. Chairman. I am just going to be
very brief.
Welcome. Thank you so much. The chairman has highlighted
very accurately some of the concerns. I look forward to hearing
your testimony today on how we can start to really make
progress. As I read the report, I am very troubled that we
continue to start and we restart and we start again, and yet
the other part that we are not doing is, when we are looking at
$3 billion in program goals, there is not really a measurable
matrix on whether we are getting there or not. And it is very
troubling when I sit there as a business guy and say, well, how
do we know if we are making great progress when we are not even
really accurately measuring it.
I represent a district that has one of the greatest data
centers in the world, Google. They know how to do it. And to
find that we have 622 human resource systems out there, another
580 financial management systems, and 777 supply chain systems,
many of which don't talk to one another, one of the advantages
of big government should be the efficiencies of systems and the
management thereof; and what we have done is we are acting like
we are a private company with thousands of different systems
that don't work.
So I look forward to hearing it. I know that we have a
Leviathan here. I mean, this is a big problem. But the other
aspect, it is very difficult for me to go back home and tell
the people why we have wasted $10 billion on terminated
projects; why we can't figure it out on a lot of these before
we spend the amount of money that we spend, before we make a
determination it is not going to work. So I would be interested
in hearing from each one of you the matrix of which we are
going to be measuring it, the goals that we are going to do,
and how we incentivize you to do that.
The chairman has held a hearing on some 13,000 Federal
buildings, and we heard that the GAO really didn't want to get
rid of them because, if the money gets sold, it doesn't go back
to the GAO. I mean, I also am finding it amazing that we have
most of our expenditures in the fourth quarter; that we have
this disproportionate amount of purchases that happens in the
last quarter of every fiscal year.
We all know what the problem is, but it really lacks
accountability. So I look forward to hearing from you on how we
are going to measure it and how we are going to fix it going
forward, and I thank the chairman. I yield back.
Mr. Mica. I thank the gentleman.
Other members? Mr. Pocan, welcome.
If no other members seek recognition, we will turn to our
panel of witnesses, and, again, we have three witnesses, and
welcome them.
First we have Mr. David Powner. He is the Director of
Information Technology Management Issues for GAO, the
Government Accountability Office; Mr. Steven VanRoekel. He is
the Acting Deputy director for Management and Federal Chief
Information Officer for the Office of Management and Budget;
and Mr. David McClure is the Associate Administrator of the
General Services Administration's Office of Citizen Services
and Innovative Technologies.
Welcome to the witnesses.
This is an investigative and oversight committee of
Congress and, in light of that, we do swear in our witnesses.
If you will stand, raise your right hand.
Do you solemnly swear or affirm that the testimony you are
about to give before this subcommittee of Congress is the whole
truth and nothing but the truth?
[Witnesses respond in the affirmative.]
Mr. Mica. Let the record reflect that the witnesses have
responded in the affirmative.
So I welcome you and we will turn to our first witness. We
will go to GAO first, Mr. David Powner. You are welcome and
recognized.
What we usually do is, it is a small panel today, but try
to keep it to five minutes. Then we will go through the three
of you and then come back for questions.
So welcome again, Mr. Powner. You are recognized.
STATEMENT OF DAVID POWNER
Mr. Powner. Chairman Mica, Ranking Member Connolly, and
members of the subcommittee, we appreciate the opportunity to
testify on the Federal Government's efforts to better manage
its annual $80 billion investment in IT. My comments will focus
on three areas: one, the Federal Government's poor track record
when it comes to delivering on large-scale IT acquisitions and
the need for greater transparency and governance; two, the
importance of the data center consolidation effort; and, three,
the need to eliminate duplicative IT spending.
Fortunately, for each of these three areas, poor delivery,
unused data center capacity, and duplication, OMB has excellent
initiatives in place. GAO's work over the years has shown that
the Government has a poor track record when it comes to
managing and delivering IT acquisitions. My written statement
lays out a comprehensive list of the many failed and troubled
projects that are highlighted here. Specifically, 15 of these
projects are examples where billions of taxpayers' dollars have
been wasted, with little to show for it.
To address this situation, OMB rolled out the Federal IT
Dashboard in 2009 to improve the transparency of approximately
700 major IT investments, and since 2010 this information has
been used to hold TechStat sessions to terminate and turn
around IT investments that are failing and not producing
results. This increased transparency has resulted in improved
governance, reduced scope, and even terminated projects.
Clearly, the Dashboard and TechStat sessions have made a
difference, but this is not enough because some agencies are
still not reporting Dashboard information accurately. In
particular, DOD is reporting no red investments. We have
highlighted DOD's inaccurate reporting at multiple hearings and
will continue to do so. When an agency is spending $34 of the
$80 billion and not reporting accurately, something needs to
change to make sure our tax dollars are being appropriately
overseen. Also, DOD has not updated most of their Dashboard CIO
ratings for about two years.
Regarding the TechState sessions, our work shows the number
of TechStats held to date is relatively small compared to the
currently reported 160 at-risk investments that total $10
million.
So for troubled projects we need more accurate transparency
on the Dashboard and even greater executive oversight.
Turning to data centers. This consolidation effort was
initiated to improve the Government's low server utilization
rates, which was estimated between 5 and 15 percent, far below
the goal of 60 to 70 percent; and this effort is to result in
$3 billion in savings. Data center closures to date and those
planned are promising. About 500 centers have been closed and
it is expected that over 800 will be closed by September. In
addition, some agencies are already reporting savings. The
Department of Agriculture recently reported that it has saved
$50 million this year, and DOD plans to save $575 million in
fiscal year 2014 alone.
Our report recently delivered for this committee showed
that OMB, GSA, and the Data Center Task Force need to step up
efforts to track cost savings and to find metrics for those
centers that remain. In fact, OMB has not been tracking cost
savings. FATAR would be extremely helpful, since it requires
the tracking and report of cost-savings and would ensure that
this important initiative would span multiple administrations.
Mr. Chairman, in addition to tracking cost-savings, there
needs to be better transparency on how many centers are out
there. When we testified before this subcommittee in May, we
reported that there were about 3100 data centers Government-
wide, only to learn from our audit work that the Government had
3,000 additional centers. Last week, OMB briefed congressional
staff that there are actually 4,000 more centers, bringing the
total to more than 7,000; and there are some fundamental
questions whether the Government really knows what it has and
why there isn't better transparency here. Timely transparency
on how many data centers are out there, closures, and cost-
savings is needed.
We also need to tackle duplication more aggressively. The
Administration's PortfolioStat process is an excellent
initiative to address this duplication. OMB states that the
Portfolio results so far have been significant and that
agencies have identified nearly 100 opportunities to
consolidate or eliminate duplicative investments that is to
result in savings of approximately $2.5 billion through 2015.
The latest PortfolioStat initiative is promising if carried
out effectively. However, I would like to make two specific
observations regarding it: savings are much higher than $2.5
billion and are more in the $5 billion to $6 billion range;
and, secondly, CIO authorities need to be strengthened at many
agencies if CIOs are to carry this out. We are currently
learning that not all CIOs have authority over commodity IT,
which is not a very high bar.
In summary, many of the initiatives over the past years
have great promise; however, each requires more leadership from
OMB and agency CIOs so that billions of taxpayers' dollars are
not wasted.
Chairman Mica, Ranking Member Connolly, this concludes my
statement. I look forward to your questions.
[Prepared statement of Mr. Powner follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Mr. Mica. Thank you.
We will go now to OMB, Mr. VanRoekel. You are recognized.
STATEMENT OF THE HONORABLE STEVEN VANROEKEL
Mr. VanRoekel. Thank you. Good morning, Chairman Mica and
members of the committee. Thank you for this opportunity to
testify on the Administration's efforts to improve the
management of Federal information technology.
Since day one, the Administration has focused on harnessing
technology to improve the operations of Government and better
serve the American people. As I saw through my nearly 20 years
in the private sector, including a stint as an assistant to
Microsoft's founder, Bill Gates, the innovative application of
technology can transform organizations, enabling them to
improve service delivery and expand customer value, while also
cutting costs. As a Federal Chief Information Officer and now
Acting Deputy Director for Management at the Office of
Management and Budget, I am charged with bringing my experience
and tools from the private sector to help take Government built
for the 20th century into the 21st.
This fundamentally requires a shift in how we think and how
we operate. As leading private sector companies do, we must
relentlessly focus on outcome and results; work collaboratively
across traditional organizational boundaries; drive innovation;
foster accountability; and, above all, put our customer, the
American people, first. High-performing organizations are
results-driven, focused on customer-facing outcomes rather than
inward-looking outputs. They set business-oriented targets,
such as revenue and profit goals, and then let operating units
determine how best to achieve them.
This thinking underpins our revamped approach to data
center consolidation. Whereas, early on we looked primarily at
raw outputs, tasking agencies with counting the number of data
centers and tallying closures, we are now building on that
first work by taking an outcome-and incentive-based approach,
focusing on optimizing total cost of ownership and efficiency
of operations instead of just the number of data centers.
Agencies are now developing metrics that drive the outcomes we
want to see: lower costs and higher productivity. Beyond
closures and savings, we are now tracking metrics, including
energy, facilities, labor, virtualization, and cost per
operating system in these data centers.
Additionally, successful enterprises are not constrained by
traditional organizational boundaries, and operate in an
integrated and unified manner. In the case of IT's, this means
treating CIO's as strategic partners, on par with the other
parts of the business. That is why, through PortfolioStat, a
data-driven review of agency IT portfolios, I am pulling
together agency leaders: deputy secretaries, CIO's, CFO's,
CAO's, chief human capital officers, and program officials. We
all sit around the table and ask them to engage beyond their
individual roles to make decisions collaboratively with a focus
on delivering value to the American people.
In taking this holistic approach to governance, we are
aiming to avoid one-off isolated actions that can lead to
unintended consequences. And by working as an integrated
leadership team, we can better identify and eliminate
duplicative and redundant investments that drain vital
resources from mission-facing programs and activities.
However, just as with any leading private sector
initiative, the success of PortfolioStat hinges on being able
to hold leadership accountable for results. This is why a key
part of the annual PortfolioStat process is following up on
commitments made in the previous cycle and evaluating agencies
on progress over the prior year, something we do quarterly.
This is also why we are sharing the results of our efforts
to generate efficiencies across the IT portfolios with Congress
through our IUIT report, the way we track cost-savings. To
date, agencies have reported over $800 million in savings from
PortfolioStat alone, with more on the way, and I am very proud
that, as of this morning, slightly over a year of having this
process in place and this cost-saving track in place, we are
announcing that we reached $1.37 billion.
But to achieve breakthrough results, we must drive
innovation throughout the Government. Innovative new
technologies such as cloud computing, open data, mobile, are
transforming how IT services are delivered and consumed.
Through these new technologies, CIOs must shift from
maintaining high-risk, high-failure monolithic systems, many of
which you have highlighted here, that sap these IT development
budgets. We need to shift to using services comprised of small,
agile modules that lower risk and support emerging needs.
In the end, all of our efforts must be in service for the
customer, the American people. The aim is to provide agencies
the ability to allocate their resources to high-value, mission-
oriented activities, rather than commodity and back office
functions, so the Government can focus on what matters most,
the citizen.
We have made significant progress over the past year, but
work is yet to be done. We must continue to engage agency
leadership through the PortfolioStat process, hold agencies
accountable for the results, track savings, and keep our sights
set on building the Government on the 21st century.
Thank you.
[Prepared statement of Mr. VanRoekel follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Mr. Mica. Thank you.
We will hear from our final witness on the panel, Mr. David
McClure, and he is with GSA. Welcome and you are recognized.
STATEMENT OF DAVID L. MCCLURE
Mr. McClure. Thank you, Chairman Mica, and thanks, members
of the subcommittee, for having me here today to talk about our
efforts in data center consolidation and our partnership with
OMB and our sister agencies to optimize new technologies being
put into the Government. I want to focus on how my office at
GSA is working in concert with OMB, the CIO Council to
strengthen Government approaches to IT management, including
consolidation and cloud security.
As Steve noted in his remarks, our present approach with
data center consolidation efforts in the Federal Government is
focused on economies of scale and achieving greater
efficiencies and overall IT portfolios by optimizing core data
center performance. This is more meaningful than a singular
focus on closure counts as the primary measure of success.
Our role at GSA is to work in concert with OMB, the CIO
Council, and its Federal Data Center Consolidation Task Force
to assemble a complete view of the data center inventory and
key variables affecting operational performance, and we do that
by providing practical tools, standard data collection
templates, guidance to the agencies for planning and executing
their strategies, and consolidating their data centers.
For example, we have created an online inventory portal
where agencies can download the data being requested for their
strategies and plans. We have developed a tool that helps
agencies identify and select their core data centers. The tool
uses nine draft criteria that we have reached consensus on
Government-wide. Most importantly, we have built a world-class
total cost of ownership model to facilitate robust data
analysis, optimization planning, and data-driven decision-
making, and it is now being used by all 24 CFO Act agencies.
Industry experts and GAO have given the model very positive
assessments. It allows agencies to analyze different scenarios
to calculate the effects of different data center optimization
strategies. It also allows, for the first time, an apples-to-
apples comparison with other agencies as they examine
outsourcing and cloud infrastructure options. Of course, we
continue to expect to enhance the model with continued input
from the CIO Council and from OMB.
We are working in concert with the community on multiple
fronts to get better results with IT investments being made in
the Federal Government. The Federal Government, for the first
time, has a comprehensive inventory of its data centers, one of
the largest cost items in the Federal IT portfolio. As a result
of our data collection and our TCO data model, agency CIO's
have more transparency into how these centers are being
utilized and viable options for optimizing their operational
performance.
In summary, the partnership that Steve and I have forged
between our offices for management change is grounded in our
own private sector experience, using industry-leading practices
that emphasize data-driven decision-making. I think Steve is
bringing the power of performance metrics to bear on the
Federal CIOs as a lever for change.
I want to thank you for having me here this morning. I am
happy to answer any questions about GSA and our role. I really
appreciate the leadership of this subcommittee and the full
committee, because it is paramount to IT reform success, and we
welcome continued interactions with you and your staff as we
find meaningful ways to facilitate effective Federal IT
investment results. And I would be happy to answer any
questions from the subcommittee.
[Prepared statement of Mr. McClure follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Mr. Mica. Okay, we will catch all of you right now, and I
will start off with some questions.
First of all, let's look, Mr. Powner, Mr. VanRoekel. We
started out in June of 2012, I guess, actually, in April of
2013 we were told that there were 3,133 data centers. Now we
are told it is over 7,000. Kind of a slight miscalculation. And
then I think there were a couple gasps from members of the
panel, joint bipartisan gasps when you said, Mr. Powner, that
the Federal Government really doesn't know what it has. What is
going on?
Let's start with Powner.
Mr. Powner. Well, first of all, the 3100, we did report
that at your May field hearing, and GAO found through our audit
follow-up, we follow up with each agency, that there were an
additional 3,000 centers that were reported.
Mr. Mica. So do you think the next time you come back we
will be up to like 10?
Mr. Powner. Well, we are up to 7,100 now.
Mr. Mica. I mean, there was a collective gasp when you said
Government maybe really doesn't know what it has.
Mr. Powner. And I stand by that statement. I think there
are 7100 now. Hopefully, we know what we have, but I wouldn't
put money on it.
Mr. Mica. And I think, actually, now, wasn't it you also, I
have your little notes here, you said the savings could exceed
$5 billion to $6 billion, rather the $3 billion that was
projected?
Mr. Powner. Well, with the PortfolioStat initiative we have
agencies reporting to us there is a $2.5 billion number
floating around out there. Most agencies reported 2.4, then DOD
came in alone and said they could save somewhere between $3
billion and $5 billion. So I think the $5 billion to $6 billion
range is very accurate for savings on PortfolioStat.
Mr. Mica. Then VanRoekel, he comes up and says today he is
proud to announce the $1.37 billion in savings. That is your
new number? Where did you get that?
Mr. VanRoekel. So, sir, we report quarterly to Congress
through a mechanism called the IUIT report. That IUIT report
cleared yesterday and was transmitted to Congress.
Mr. Mica. For today? So instead of, where the hell were we
here? We were at, this is projected savings were about $300
million by OMB according to a May report, and now you are
saying we have jumped a billion?
Mr. VanRoekel. There are sort of two numbers that we are
tracking simultaneously here.
Mr. Mica. Is that from the closure of the 484? It says you
have a total plan closure of almost double that, 855. So have
you had a sudden lightening expansion.
Mr. VanRoekel. You are actually looking at two separate
tracks of numbers. One number is just data. The number you are
quoting with the $300 is just data centers.
Mr. Mica. Okay.
Mr. VanRoekel. There is a broader set of IT savings which
have data center implications.
Mr. Mica. So this is sort of a self-reporting and they are
telling you that they have saved money by doing that.
Mr. VanRoekel. No. We actually track at the investment
level through the IUIT report, and I would encourage you to
check it you.
Mr. Mica. So it wouldn't be reflected in this. Then this
$300 still stands, $300 million savings OMB as of May?
Mr. VanRoekel. In May, yes. In May, yes.
Mr. Mica. So this is another billion. And where would you
attributed that billion coming from?
Mr. VanRoekel. I am happy to share, with the permission of
the Appropriations Committee, who we submit this report to, the
report.
Mr. Mica. Some data?
I think it would be good for us to know. If you are making
progress and it can be documented, we would like to know about
it.
Again, from the consolidation standpoint, we are somewhat
still stuck in neutral, Mr. Powner? Actually, it looks like we
are not only in neutral; it looks like we have gone in reverse
because we actually have found more of these non-core centers
than we anticipated. So we are behind the eight ball a bit.
Mr. Powner. On the data center consolidation, that now has
been merged with PortfolioStat, but to be real clear on data
center consolidation, there is to be 800 centers closed, more
than 800 centers closed by the end of this fiscal year, by
September.
Mr. Mica. What happened is we start out, even if we close
the 800, that would have been 800 out of 3,100, and now we are
looking at 7,100 or more, because we are not sure exactly. So
it doesn't appear that we are making progress. If it was 800 of
3,000, that would be somewhat more significant.
Mr. Powner. Mr. Chairman, I think there are fundamental
questions about how many we will ultimately close and what the
savings will be, and we agree that we need to optimize what
remains, but still, with that large number of closures, there
should be associated savings with that.
Mr. Mica. Well, again, it sounds like you have testified,
too, that there is even more potential savings, so we would
hope that can be, the problem seems to be, VanRoekel has great
credentials for his position, coming from the private sector
and observing sort of the Federal mess and trying to straighten
it out, but we have tried several tools that don't seem to be
that effective. We had this data consolidation, whatever the
FDCCI, and then we have gone to this Pro whatever it is, and it
doesn't appear that these tools are that effective.
Okay, we are developing matrix. You said we are continuing
to developing matrix to, again, get a handle on this or to move
forward with realizing these savings and consolidation. But is
this an evolution of the two or three sort of programs you have
already set up? Are they an evolution? Explain what you are
doing to get these agencies to maximize the potential for
consolidation and savings.
Mr. VanRoekel. I think the key is every day in the private
sector you think about what impact are you driving, what is the
result you are trying to affect. It could be selling products;
it would be moving the stock price; it could be creating value.
And thinking about that impact in working your way back from
that is key.
So what we hope for here is definitely cost-savings and
closures. That is an ends, not a means. And the means to get to
that is thinking about what applications am I running; how many
email systems do I have; what kind of things am I deploying;
what are the characteristics of the outcome I want to drive. So
if you ask any CEO their advice, they will say measure what you
care about, and what I care about is the optimization.
Mr. Mica. That is the end results in providing the IT
services, but when you have, okay, let's go back to Powner is
telling me, what did you testify, server use is 5 to 10
percent?
Mr. Powner. Well, when we started the data center
consolidation initiative, it was because average server
utilization across the Federal Government was 5 to 15 percent.
So that is a lot of unused capacity.
Mr. Mica. Exactly. Exactly. That would be one of the sort
of fundamental things I have. Setting up servers, the energy
they use, the space they use, the inefficiency of 5. Where are
we now? VanRoekel, it seems to me like you would look at that
to begin with. I mean, getting the mission done is good, but
you have people, they can certainly do the mission with lots of
servers and all of these centers, but somehow the core expense
and the consolidation efficiency goes back to the equipment,
the hardware and the utilization of software that we aren't--
Powner, do you continue to monitor this? Where are we at now
with, say, the utilization of servers?
Mr. Powner. We are not monitoring that now. We were
monitoring that at one time because that was stated in agency
consolidation plans, which we don't currently have visibility
into that. I think a key question is this: The core centers
that remain, so we are going to close several non-core centers.
The core centers that remain, where are we at on utilization on
those. That ought to be a key metric.
I agree with Steve, and there are a lot of metrics on power
usage, efficiency, and those types of things, but if we are
attempting to solve a problem of low server utilization, the
core ones that remain, did we fix the problem. And we are ready
to go out and measure that once those core centers are up.
Mr. Mica. But, to me, that would be sort of fundamental.
You are probably not going to do away with the core centers,
but better utilization.
Mr. Powner. Absolutely.
Mr. Mica. And if you are not doing that and you want to
accomplish a mission, you look at the alternative. If it isn't
using that core, going to cloud or some other more efficient
thing. But the problem is you go back to the matrix. We have
had a couple of plans to kind of get this thing going and to
try to move the consolidation, better utilization of the assets
that we have, but that hasn't, apparently, worked.
Now, we will probably be here in six months and haul
everybody back, and we may get more discovery of non-core data
centers and maybe a few more of those closed, but it doesn't
appear like we are really getting a good handle on a means to
make this happen.
And then you have McClure here, who comes from GSA.
McClure, you guys haven't set a very good example in your own
operations. You have closed only one, less than one percent.
You have 115 non-core data centers. What about yourself not
operation? Anything to say on that?
Mr. McClure. Well, I think that data that you are looking
at was the second quarter reporting from GSA. Since then there
has been a lot of movement. GSA owns, primarily, non-core data
centers, small data centers, and we were in the middle of
collecting the date when that second quarter----
Mr. Mica. So what is the good news?
Mr. McClure. Well, the good news is we expect to close 75
of our 116 non-core data centers.
Mr. Mica. Expect to, but how many?
Mr. McClure. Thirty-eight will be done by the end of fiscal
year 2013; another 37 in fiscal year 2014.
Mr. Mica. So we will run from one to add 37 more is what
you are telling us now.
Okay. And, again, you have cited some of the things you are
trying to do. Let me look at my notes here. You ever listen to
WTOP in the morning? They have all these guys advertising that
they can do this. That is the lighter side. I listen to WTOP in
the morning and I swear they have 10 ads, by the time I am
through shaving, on who can consolidate IT centers and how they
do such a great job. It doesn't sound like anyone is listening,
though.
But you are still developing tools to evaluate all of this
and you have put some tool, the data center optimization.
Explain that and how it works.
Mr. McClure. Well, I think it gets at the heart of what
Dave was saying and what Steve was saying, and what you are
saying, Mr. Chairman. We have to have a data-driven decision-
making process and transparency into that data. The GSA, our
office has created a total cost of ownership model which would
provide data into a lot of these areas that we are have been
talking about this morning; power capacity, utilization.
Mr. Mica. But you are laying out sort of the things that
you could do; it is not things that you should or must do?
Mr. McClure. No, no, this model exists. It is in the hands
of all the agencies. It is an optimization planning tool that
they can look at.
Mr. Mica. Again, part of the solution would be to, and I
have gotten to this. I have run out of time, but is to make
certain, maybe VanRoekel has the tools to do this. Maybe from a
procurement standpoint you would dictate or mandate that such-
and-such has to be achieved. VanRoekel has to go in and sort of
force them. I have always said we give you budget authority,
and they aren't performing and you just, you know, you have to
have some teeth in the process to get people to do things.
But I will leave it at that. I will come back. I want to
give others an opportunity.
Mr. Pocan.
Mr. Pocan. Thank you, Mr. Chairman.
This is a question maybe for Mr. McClure. According to the
GSA website, the goals of the FedRAMP program include
increasing the confidence and security of cloud solutions and
security assessments. Under the process, cloud service
providers must adhere to certain security requirements and
undergo an independent audit.
I am wondering a little bit about the confidence that
agencies like DHS and DOD and NSA can have with information
that is stored on the cloud, that it won't be compromised in
some way.
Mr. McClure. It is a great question. I think if we follow
the baseline security requirements that we have established in
FedRAMP and we do the independent certifications by these
third-party assessment organizations that know how to assess
cloud security, I think we will have much, much greater
confidence that our data at all trust levels is really secure.
I think FedRAMP is trying to instill in the Government a trust
level that has not been there before, and to date we are
finding it really does improve the confidence and trust of both
the agencies and the providers that they are doing that.
Mr. Pocan. And given all the controversy about cyber
attacks and things, what confidence do we have that, as hackers
are constantly revising how they are finding ways in, how
confident are you that that FedRAMP process will provide enough
flexibility that we can keep up with any evolving schemes or
new technologies to try to break through?
Mr. McClure. I think it will do it in two ways: number one,
we are always going to be changing the FedRAMP security
program. As the NIST security guidelines change, we incorporate
that into the FedRAMP baseline, so it is always going to be up
to speed with what NIST is recommending the Federal agencies
do.
Secondly, we are putting a lot of stock in the ability of
these agencies to do continuous monitoring so that they have a
real-time operational view of their security posture of
themselves and their cloud providers. That is the best defense
we can put in place, is having very robust, continuous
monitoring.
Mr. Pocan. And these audits, who is going to do the audits
and what do they typically consist of?
Mr. McClure. The audits that are done for FedRAMP
certification are done by independent assessment organizations
either off a list that we have put together, where companies
have passed our accreditation that they actually have the
capability to do the cloud assessments, or independent ones
that exist in the marketplace. Based upon those audits, we can
then review that in my program management office and determine
whether there are any follow-up questions. Or the agencies can
use them and feel much more confident that the audit has been
done consistently, according to baseline standards, and is
repeatable, can be reused across Government.
Mr. Pocan. And the timeline for FedRAMP to be fully
operational?
Mr. McClure. It is moving into fully operational status
now. We have run it for a year in what we call an interim
operating capability. We didn't want to roll this out
Government-wide until, if you pardon the expression, kick the
tires, so we wanted to make sure this worked. We wanted an
opportunity to refine it for success, and we think we are ready
now.
Mr. Pocan. Thank you, Mr. Chairman. I yield back the
remainder of my time.
Mr. Mica. I thank the gentleman.
Mr. Meadows, the gentleman from North Carolina.
Mr. Meadows. Thank you. I appreciate the insight. I must
say that I am troubled by some of the things that I am hearing
here.
Mr. McClure, let me start with you. So, to date, you have
closed one data center, is that correct?
Mr. McClure. I will check as of today. I don't know if it
is one or more, but there are several that we are nearing the
close.
Mr. Meadows. My data says that you have closed one in three
years. And you are going to now, according to your testimony,
you are going to close 37 in two months? What happened to go
from one in three years to 37 in the next two months? How do
you plan to accomplish that?
Mr. McClure. Well, I don't operationally own the task
myself, the CIO of GSA does.
Mr. Meadows. So if it doesn't happen you can blame them?
Mr. McClure. I am not blaming, I am just stating a matter
of fact.
Mr. Meadows. Well, we need to know where does the buck
stop. Does it stop with you or does it stop with GAO?
Mr. McClure. It should stop with the head of the agency, as
with all these matters.
Mr. Meadows. So it stops with you.
Mr. McClure. I am not the head of the agency; I am the head
of an office that provides the tools that we talked about to
help get this job done.
Mr. Meadows. So if we bring you back in two months and 37
of them are not shut down, who should we hold accountable?
Mr. McClure. Well, the CIO holds responsibility in each
agency for doing the data center consolidation work and for
estimating plan closures, so the CIO owns the issue. The head
of the agency is ultimately responsible. So those are the two
individuals at GSA that have their eyeballs on the situation.
Mr. Meadows. So what happens if they don't make the 37?
Mr. McClure. I think you should ask for an explanation.
Mr. Meadows. So Ms. Coleman would need to come here and say
why she couldn't get it done?
Mr. McClure. I would recommend that, yes.
Mr. Meadows. All right.
Mr. McClure. I think that would be true, by the way, across
the board for every agency. A lot of the responsibility here
lies at the CIO and head of the agency level. So the same
conversation could be repeated across multiple agencies.
Mr. Meadows. So essentially we have a whole bunch of people
that come and give testimony before Congress, but really don't
have the responsibility for implementing those things. So we
have hearing after hearing after hearing and nothing gets done?
Is that what happens?
I mean, I am at a loss on how one got closed. I think you
have three core centers, 115 non-core centers, and we have one
closure in the last three years, and now, all of a sudden, we
are going to ramp up. Why is that? I mean, who made that
decision to ramp up and how is that going to happen?
Mr. McClure. Well, I think the thing to remember is that,
again, we had a definition change. That is the reason why the
number of data center number changes.
Mr. Meadows. Okay, but we have heard from testimony, I
think over here, that actually the number of data centers have
increased. But they really haven't, so they have stayed static.
I think OMB, according to your briefing, you knew as far
back as June of 2012 that we actually had 6700 centers, and
that has grown to 7100 now, is that correct?
Mr. VanRoekel. Sir, I actually evolved and put more rigor
behind the definition to expand the definition.
Mr. Meadows. Of what a data center is?
Mr. VanRoekel. Of what a data center is. What I didn't want
to do is have Federal agencies either splitting up big ones and
putting them in small rooms or hiding computing resources,
because you add up those small ones, you are going to get as
much as a big one, and I want to track all of them. Our prior
definition only tracked the very large. If we change the
definition----
Mr. Meadows. So your prior definition of large ones----
Mr. VanRoekel. Was over 500 square feet, I believe.
Mr. Meadows. Okay. And that was 6700?
Mr. VanRoekel. Was the 3,000.
Mr. Meadows. Is the 3,000 number that they originally
reported.
Mr. VanRoekel. So I came in and said I don't want anything
hidden, I don't want resources in dark corners.
Mr. Meadows. Okay, that is reasonable.
Mr. VanRoekel. I expanded the definition to say we are
going to go into the small ones too. That contributed, then, to
the----
Mr. Meadows. Okay. So what you are saying is that expansion
is really a function of changing the criteria of what we call a
data center.
Mr. VanRoekel. To be more comprehensive and derive better
inventory.
Mr. Meadows. All right, so let me pick up on one other line
of questioning, then, because I think what we are saying is our
server capacity right now, according to Mr. Powner, we are
operating at 5 to 15 percent. Would you agree with that?
Mr. VanRoekel. I don't have raw data, but that is probably
an industry----
Mr. Meadows. So why are we continuing to buy new servers,
then? And I am not talking about replacing existing servers. We
are actually purchasing additional servers. So why are we doing
that if we have excess capacity?
Mr. VanRoekel. The way that technology grew up and the way
that we built data centers very much follows what the private
sector did, which it was----
Mr. Meadows. Well, if I did this in the private sector I
would go out of business.
Mr. VanRoekel. Until about mid-2000, the private sector did
this.
Mr. Meadows. I understand.
Mr. VanRoekel. Server utilization for corporate customers--
I was part of the leadership to the server division of
Microsoft--was 7 percent utilization prior to this new
technology coming onboard called virtualization that allows you
to put more stuff on physical hardware. You need new hardware
to run that technology. There is new investment.
Mr. Meadows. But that is what I am saying, this is not
replacement of existing servers; these are actually additional
servers, according to the data that I have.
Is that correct, Mr. Powner?
Mr. Powner. The increased inventory is existing, what
currently exists at these agencies, where we are discovering
more based on the new definition that Steve laid out. So----
Mr. Meadows. So this is just a definitional change, it is
not an actual increase in the number? Are you all sure about
that?
Mr. Powner. I think it is both. I think it is both. Here is
what happened. Initially we started off, the definition was 500
square feet or larger. Then we expanded the definition and
things were smaller. Steve did the absolute right thing because
there were a lot of opportunities to save by including these
additional things. So the definitional thing resulted in an
uptick.
But I think over time, based on our audit work, because we
go in and ask agencies what their inventories are, we see these
inventories continuing to grow over time. So it is twofold, it
is the increase, it is the change in definition, and it is also
they are discovering more what they have.
Mr. Meadows. All right. So let me go back and pick up on
the national security. If we have 7100 data centers, from a
security standpoint how do we manage that security aspect from
the standpoint of cybersecurity, a number of things in terms of
attack? Because it seems like it would be the more concentrated
it is, the easier it is to provide a higher level of security.
So, as we grow that out, is there a matrix right now that you
are looking at to try to say, well, optimum efficiency would be
to get down to 2,000 servers? I mean, do you have a number?
Mr. VanRoekel. It is less about the raw, the number of data
centers, of course, is important, as we are talking about
today. Cybersecurity related to data centers is just a little
bit different in the way you think about that. And the way we
build our cyber capabilities is grounded in FISMA, the Federal
Information Security Management Act, and we use a process
called CyberStat, much like PortfolioStat, that tracks key
metrics, and effectively what you want to make sure you are
doing is that when traffic comes and goes from these centers,
that they are going through these trusted Internet connections
and that we have capabilities there----
Mr. Meadows. But fewer data centers would make that an
easier task or not?
Mr. VanRoekel. I think it would make it an easier task,
definitely, and that is a byproduct of----
Mr. Meadows. So where do we need to be? If you had control
over all the other agencies and you are the guy that is in
charge and ultimately the buck stopped with you, how many data
centers do we end up? What would you do if you were a private
sector guy trying to make money here? What would you do?
Mr. VanRoekel. Fewer is better and optimized is better.
Mr. Meadows. How many fewer?
Mr. VanRoekel. It would depend on the size of the agency,
the mission of the agency, what is the relationship of that.
Homeland Security, for example, has defined three core data
centers.
Mr. Meadows. Sure.
Mr. VanRoekel. I think that is a good target to think
about. And these are highly optimized, they follow all the
guidelines that we have put forth in this approach to say where
we need to go. I can't extrapolate for the whole of Government.
Mr. Meadows. And that is fine, but so we don't just have
hearing after hearing, I would ask if you would go ahead and
try to put together what the plan is so we know whether--one of
the criticisms of your agency, and I understand this is a new
role, is that you are not measuring, that you are not
effectively measuring, and I think that that can be maybe
accurate in some ways and inaccurate in others. I will give you
that. What I would love to see is how do we know whether we are
being successful or not, so really would love for you to follow
up on that, have your staff follow up on that.
The second part of this, and this is probably as critical,
what can we do as Congress to give a real incentive for us to
save money. I have hearing after hearing where we save billions
and billions of dollars, but yet we ask for more and more
money. So what I would love to do is to find a real way where
you say, Mr. Meadows, if we did this, if you offered us this,
we could assure that we could get $50 billion in savings or
over the life, or whatever it is.
But I am talking about real incentives where, hopefully in
a bipartisan way, we can come up with something that gives you
incentive. I would love to hear that from each one of you, not
in terms of answering that question, but as a follow-up, if you
would submit that to us.
Thank you so much. I yield back, Mr. Chairman.
Mr. Mica. I thank the gentleman from North Carolina and
recognize our ranking member, Mr. Connolly now.
Mr. Connolly. Thank you, Mr. Chairman.
Forgive me, Mr. VanRoekel and Mr. McClure, for not being
here for your testimony, but I have been trying to catch up. I
had to move an amendment on the Floor and the Republican
manager accepted my amendment. So I don't know whether it means
it was just awful and brilliant or whether apparently I wrote
an amendment that was so weak that even for them it was
acceptable. I don't know, but I will take the gift.
Mr. Powner, if I understood your testimony correctly, you,
inter alia, said we seem to be sliding backwards in terms of
certain metrics with respect to, for example, data center and
Dashboard, is that correct?
Mr. Powner. I think in terms of the Dashboard, yes, on the
data centers what we want to do is we want to track savings. I
think we want to know what the inventory is, transparency on
what the inventory is, what the savings are, and then how we
optimize what remains. And transparency around that is key.
Mr. Connolly. Okay. I also thought I heard you say there is
some concern that the Government doesn't fully know what it has
when it comes to data centers.
Mr. Powner. Well, when we see the inventories growing as
they have over time, you can just take the snapshots in time;
it keeps growing, so there is still a concern have we still
captured everything.
Mr. Connolly. And I also thought I heard you say that you
thought the FATAR legislation passed out of this committee and
passed on the floor of the House would be actually helpful to
the Government in trying to get its arms around this subject
matter.
Mr. Powner. On data center consolidation optimization, it
would codify that in law, clearly.
Mr. Connolly. Mr. VanRoekel, your reaction to those
elements of Mr. Powner's testimony?
Mr. VanRoekel. As I mentioned earlier, I do agree that
there is a level at which the inventory management, when you
cast a net across the Federal Government, there will be things
you probably aren't finding just given the sheer size. I think
our rigor around the definition and modifying that definition
over time to capture more of what is out there has been the
driving force behind the number increasing, less about sort of
inventory management. But I do think there will be edge cases
where we will have more come up in different cases.
Mr. Connolly. Can I interrupt you there? I take that point,
but if you look at GAO's report agency-by-agency how you are
doing in data center consolidation even before the announcement
of an additional 4,000, it is pretty slim pickings for most
Federal agencies.
Mr. VanRoekel. If you visit a data center and understand
how a data center works, it is not literally walk up and pull
the plug and say I am going to shut this thing down. And the
incentive structure you want to establish here needs to be one
where you drive to optimization, because if you lay a metric
out and say close data centers, cut them by 30 percent, cut
them by 40 percent off just a denominator of inventory, what
you will have happen is Federal agencies, in many cases, will
literally take a forklift, pick up small data centers, move
them to a larger room and plug them back in.
Mr. Connolly. And that is called compliance.
Mr. VanRoekel. And that is called compliance. So you have
two next to each other and they remove the wall between them,
thus reducing that by 50 percent. That doesn't get where we
want to go, which is to get that utilization up, get the
optimization up, get the service out of that data center to
increase in such a way that you drive better outcomes for the
mission of that agency and for the American people.
So the incentive structure has to match to the measurement
in a way that I think needs to drive the behavior we want to
see and the outcomes we want to drive. So the uncertain budget
environment we have been in for a while, CIO authorities, which
the FATAR bill looks at that, and other driving factors
contribute to this phenomenon of not being able to just shut
one off.
Mr. Connolly. I do want to be clear. The FATAR legislation
is not sort of a Luddite approached to the subject matter; it
actually does require tracking and it does have other
measurements about utilization and so forth that are
incorporated into the concept. So I think we kind of took your
point. But I would also say to you that the metrics of the
number of data centers and the need to reduce them came out of
this Administration. I mean, that wasn't something that ended
up here; that was something that came out of the White House.
Mr. VanRoekel. And we still stand behind that direction.
Mr. Connolly. We kind of hope you do. We take your point,
and we are not always good about nuance, but we are going to
try to be responsive on a nuance way, the bottom line is
efficiency, utilization. But we have to have some metric that
says we have too many of these things, and that alone tells us
we are inefficient; that tells us, you know, sort of there's no
place like home syndrome: I know Steve has a better one, but
kind of like ours right here, and we are not going to give it
up willingly, and we are going to use every bureaucratic trick
in the book to protect and preserve it, irrespective of
utilization. Utilization, we can have Harry and Shirley go
there twice a day instead of once a day and get up to
utilization; it doesn't really get to what we are trying to get
at.
And I was saying to the chairman, he shares my view that I
think the fact that you are now also at OMB gives us some hope
that from a management point of view we can perhaps persuade
people that there is a better way of doing this and it is win,
win, win. We can save on energy, we can save on budgets,
especially in a time of contracting budgets, and we can make
ourselves more effective. So we want to be supportive of that,
but we are frustrated that the numbers are not particularly
felicitous. And I hope you can understand that, from our point
of view, that is to say.
Mr. VanRoekel. Yes, I can definitely understand that.
Mr. Connolly. Okay. The IT Dashboard, the Department of
Defense reports zero investments with significant concerns and
has not updated the status of most of its investment over two
years, and that more TechStat reviews obviously need to occur,
as Mr. Powner said in his report. What are we doing about that?
I mean, that is actually, to me, astounding. Here is the
biggest expenditure of Federal dollars, here is the biggest
investor of Federal dollars, here is the biggest client we
have, from your point of view, and it hasn't even essentially
updated its Dashboard in two years. Huh. I guess we have been
busy doing other things.
Mr. VanRoekel. I think it speaks to a couple aspects. One
is we don't, in OMB, in my office, track the self-reported
status as the key indicator of performance on investments. It
is a fool's errand to track a self-reporter. You would never
have a contractor self-report their results or things like
that. So we go deeper than that and look at how often do they
change schedules; where are they on budget, are they hitting
budget; what is their time to delivery on services, all of
these kinds of things.
I have actually added features to the IT Dashboard in the
last couple years that give me indications when agencies go in
and do re-baselines, meaning they have changed their data in
some fundamental way. Those, to me, are the red flags you want
to look at. And where we will lift up and say there is
something going on there, where self-assessment will never do
that for you. And I agree that it is laughable, to some degree,
on DOD not reporting any core investments, but we know; we
track and we know where those are.
The second part of it is CIO authorities. And I think
looking at the authority of the CIO, the person whose picture
is next to all those investments is Terry Takai. Terry Takai
has very little influence over most of the investments that you
are looking at on that Dashboard and has very much an
inability--you know, she is reporting what she gets from the
self-reported aspects of the people out in the periphery, and I
think it speaks to a larger theme of something we need to look
at in Government around what is the authority of the CIO or
whose picture should be next to that.
Mr. Connolly. Well, funny you should bring up that subject,
Mr. VanRoekel. FATAR addresses that issue in terms of the
streamlining of CIOs in Federal agencies and the infusion of
authority, responsibility, accountability in a CIO, a principal
CIO for each of the 26 major agencies. And I think Mr. Powner,
in previous testimony, has highlighted that as well. So I
commend the bill again to you, because I think it tries to move
us in that direction without a heavy hand. But it is trying to
inculcate more flexibility for a chief CIO. It doesn't abolish
all other CIOs, but I would commend it to you that it is
designed, again, to address the very thing you are talking
about.
Now, DOD, because of jurisdictional issues, is not directly
addressed, but sooner or later it will be, and we will be glad
to work with the Administration to make that happen.
Some questions have come up recently about energy savings
performance contracts, and let me ask what is the role of OMB
in approving such projects.
Mr. VanRoekel. OMB doesn't, accept for interpretation of
policy or matters where procurements reach a certain threshold
where we have a review board process that is part of our normal
Office of Federal Procurement Policy work, review those and
provide counsel to agencies; it is the agency's decision to go
forward and the Procurement Office to go forward with the
procurement.
Mr. Connolly. So OMB is not going to play any kind of
direct role in the awarding of such contracts, the approval of
such contracts, the extension of such contracts, or even just
conceptually the general approbation of or disapproval of those
as a tool.
Mr. VanRoekel. We provide guidance in that context.
Mr. Connolly. Okay. I thank you.
And, Mr. McClure, when we had our field hearing there were
concerns raised about GSA's performance with respect to data
center consolidation. You were not there to answer those
questions, so I want to give you, as my last question before we
have to vote, an opportunity. How is it going? I think after
the hearing you did have some announcements, and that was good.
It would have been nice to have them at the hearing. But the
concern is that this is a sustained trajectory, not just a
let's please them and give them something to report, and it
kind of looked like the latter more than the former. So
reassure us that that is not true, please.
Mr. McClure. Well, I do want to reassure you. I think it
might have the appearance of that, but, as I was explaining
while you were gone, there was a lot of data collection going
on that had not ended by the second quarter reporting period,
which is the data that the committee had at that time. So, as a
result, shortly after the hearing, with updated numbers, our
situation looks much better; 38 planned closures by the end of
fiscal year 2013 and an additional 37 in fiscal year 2014.
Mr. Connolly. Can I interrupt you there, Mr. McClure, just
to satisfy myself? The chairman was reminding me the report we
had was you have done one. Now you have 37 more in the
pipeline, but how could we have taken so long to just have one?
Mr. McClure. Well, I will give you some explanation that
revolves around what Steve just said, and that is that CIOs in
the Federal Government often don't have complete control over
all data centers in that department or agency. That was the
case until May of this year, June of this year, in which
Administrator Tangherlini consolidated CIO authority under a
single CIO. So the ability to collect this data I think has
been greatly enhanced with that kind of authority being vested
in the CIO.
Mr. Connolly. So, in other words, for example, had the
FATAR bill been law, we might have been able to have happier
numbers much sooner.
Mr. McClure. I think it can help, because it has helped
there.
Mr. Connolly. Stop right there; you are doing fine.
[Laughter.]
Mr. Connolly. Okay, we have to go vote. This is part of a
dialogue. I certainly appreciate you being here. I hope we have
conveyed, through our frustrations we share on a bipartisan
basis. We also, on a bipartisan basis, want to be partners. We
want to help. This is an important part of Government that
doesn't get sufficient attention. But in terms of our future,
the investments we make in technology are going to drive
everything, not just something. And from megadata evaluation to
cybersecurity attacks to wonderful ability to do great things
more efficiently, you all have the keys in your hands to help
us make that happen and we want you to be successful.
So we want to try to help create an environment for
success, so I hope you look at it in that spirit. Congress has
an oversight role and we have to throw the flag down when we
think something has gone wrong, but that is not the end game
for us. And I can say my partner here, the chairman, Mr. Mica,
has gone out of his way to try to create this subcommittee as a
forum to be helpful and to be useful. So we look forward to
working with you and thank you so much for being here today.
Thank you, Mr. Chairman, for the hearing.
Mr. Mica. Thank you, Mr. Connolly.
I will re-echo his comments that we are trying to find some
constructive means of helping you move forward, both GSA has an
important role, certainly OMB, and we want to talk some more,
Mr. VanRoekel. I want to also see if the legislation we have
pending provides you with the tools. You need some teeth. You
are a nice guy. You came from the private sector and mission-
oriented.
I love all that, but the difference is in the private
sector the people that are involved are business people who are
bottom-line people; they are doing everything they can to bring
the cost down, the efficiencies, maximize the assets that they
have. Here you have a Federal agency; it is just more manna
from the Treasury and they don't have that same incentive. But
we are going to figure out a way to give you all the tools you
need to help us get the job done.
And thanks, Mr. Powner. Keep us posted as you reveal more
of the findings that cause a bipartisan gasp in any panel of
Congress like you did today.
So, with that, I am going to ask that we leave the record
open for two weeks. We have additional questions we didn't get
to. We want to have those answered in the record. So, without
objection, so ordered.
There being no further business before the Government
Operations Subcommittee, this hearing is adjourned. Thank you.
[Whereupon, at 10:53 a.m., the subcommittee was adjourned.]
APPENDIX
----------
Material Submitted for the Hearing Record
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]