Transcript

Department of Defense Enterprise Cloud and its Importance to the Warfighter Media Roundtable

Aug. 9, 2019
Department of Defense Chief Information Officer Dana Deasy and Director, Joint Artificial Intelligence Center Lieutenant General Jack Shanahan

MR. DEASY:  Good morning.  

So as Elissa said, I’m Dana Deasy.  So I have overall responsibility for the DOD CIO [Chief Information Officer] activities across the Department of Defense as, I think, some of you have written in the past.  One of the major programs I lead is the digital modernization for the overall Department of Defense, which has four key components to it.  One of those is cloud; the others being A.I. [Artificial Intelligence] -- which is why Gen. Shanahan's here with me -- as well as next generation command control communications, as well as cyber.  I always like to point that out because our cloud strategy sits as part of much broader digital modernization program.  

We wanted to bring you all together today to give you an update on where we are at on the overall program, and have a couple of very specific conversations today, the first one of which is the recent announcement that has come out in the press regarding Secretary Esper.  As you know, he assumed his role as secretary of Department of Defense approximately three weeks ago, and you have all heard that in assuming that role, one of the things he has said is that he is going to personally take a hard look at the overall JEDI [Joint Enterprise Defense Infrastructure] program. 

The other thing I want to -- I'll talk more about that in a minute.  The other one is there's a lot of narratives going on, on JEDI, and what we wanted to do today was kind of set the record straight on some of these narratives.  There is a couple of really key narratives in there, and I'm going to walk you through those narratives, and then that's going to get to the conversation on the warfighter.

So first on Secretary Esper.  So Secretary Esper made a commitment during his confirmation hearing that he would take a look at the overall JEDI program.  People have picked up this word, quote, "pause".  There is not a pause on the overall JEDI program, meaning that we are still a number of weeks away from completion of the overall evaluation.  Now, some of you are going to ask about the fact that I had stated earlier that we were looking towards a date in August, and I've always used the expression "not earlier than August".  

Because a lot of our attention and time has been placed on the Federal Claims Court, which you know we came out with a very positive ruling on that.  We are now able to pivot our full attention, the energy of everybody's time and efforts back towards the completion of the evaluation.  It's a highly-technical evaluation at this point that we are going through with the finals.  That still is going to take a number of weeks to complete, so therefore, the continuation of the JEDI evaluation work has not been paused.  That work continues on.

What is occurring is Secretary Esper has asked to go through a series of sessions to fully comprehend the overall JEDI program.  I want to give you some real specifics as to what's going to be taking place over the next several weeks.

So we have put together a series of education programs that allow him to get deep understanding.  As you can imagine, when he was secretary of Army, in that role he was aware of the JEDI program, but in that capacity he did not have to have a complete, deep understanding.  In his new role, given that this is a large acquisition program, you would expect any new secretary taking over the DOD to want to have a -- a strong understanding of any major acquisition program.  So that's exactly what we're doing here. 

So some of the things we're going to be discussing with him over the next several weeks is why does the DOD need an enterprise cloud?  We're going to get into a deep conversation of what are the activities we're trying to solve for, what is the urgency around this, how to support the warfighter.  

We're going to get more into that conversation with Gen. Shanahan today.  What are the security implications, how does this help the security standpoint, what are the securities from a commercial cloud standpoint, et cetera.  So we're going to have a session around that. 

Then we're going to follow on sessions around, how does JEDI fit into a much larger overall Department of -- Department of Defense enterprise cloud strategy?  That is a mouthful.  So we're going to explain to him, what is the overall strategy for the Department of Defense when it comes to enterprise cloud?  Where does JEDI fit into the much larger strategy?  And we're going to have a bunch of questions I'm sure we'll go back and forth with on that.  

And then finally, we're going to have a deep set of discussions on the JEDI program itself.  How was it created, who were all the people involved, what were the things that we needed to consider bringing into an RFP [request for proposal] process, how was that constructed, what's the selection criteria. 

And you can imagine, there will be probably a series of follow-on questions and actions out of that.  This is going to be an iterative process.  You're going to see, we're going to start very high and then we're going to go into a lot of very specifics.  

He needs this because he obviously has a role, to weigh into the overall direction of this program.  And for him to be able to do that, he needs to first go through a series of deep education sessions.  So that's what we'll be doing with him over the next number of weeks.  While, in parallel, we are still continuing through the technical evaluation process as part of our overall source selection. 

Two, out on the DOD website, we put out what is called the Enterprise Cloud Fact Sheet.  I think this went up sometime here in the last 24 hours, that's available to all of you. 

And what -- why have we done that?  So there has been, I mentioned earlier, a number of narratives that have been picked up in the media, about the cloud program.  What we've tried to do is literally take those narratives, and have put facts behind them. 
Let me just give you one such example.  One of the narratives says that the DOD is doing a single cloud, that we're going to pick a vendor that's going to run all the DOD cloud.  It's going to receive a $10 billion award and they're getting a 10-year contract.  So that's a narrative that's out there. 

The fact is, we already, today, have a number of clouds.  We're already spending well in excess of half a billion dollars annually on a multiple number of clouds inside the Department of Defense.  

The winner of this actually receives $1 million, not $10 billion.  It is a two-year contract with a three-year --- three-year plus two-year extension.  And so if you were to execute all the extensions, and if we were to pivot workloads, we believe it could, over a 10-year period, generate up to $10 billion.  

But in no way is this is a $10 billion cloud.  It is technically a $1 million award with a two-year contract with extensions that follow on.  That will be complementary to other cloud awards that we will continue to do across the Department of Defense. 

Finally, the last narrative that continues to get a lot of play is, this is not in the best interest of the warfighter.  So today, what we want to cover is setting the record straight, that the warfighter is absolutely waiting for this.  This is an imperative to what they need, each and every day, to defend and to execute their missions. 

And to that end, I've asked Gen. Shanahan to join today.  As you know, Gen. Shanahan heads up A.I. for the Department of Defense, the Joint Artificial Intelligence Center.  What he's going to do is, spend some time walking you through why this matters to the warfighter. 

LIEUTENANT GENERAL JACK SHANAHAN:  Well, thanks.  

Good morning to everybody, for those who don't already know me.  I'm Lt. Gen. Jack Shanahan, the director of the Joint A.I. Center, or the JAIC.  And before that, I was -- served as the director of Project Maven under the under secretary of defense for intelligence. 

The warfighter needed enterprise cloud yesterday.  Dominance in A.I. is not a question of software engineering.  But instead, it's the result of combining capabilities at multiple levels: code, data, compute and continuous integration and continuous delivery.  All of these require the provisioning of hyper-scale commercial cloud. 

For A.I. across DOD, enterprise cloud is existential.  Without enterprise cloud, there is no A.I. at scale.  A.I. will remain a series of small-scale stovepipe projects with little to no means to make A.I. available or useful to warfighters.  That is, it will be too hard to develop, secure, update and use in the field. 

JEDI will provide on-demand, elastic compute at scale, data at scale, substantial network and transport advantages, DevOps and a secure operating environment at all classification levels.

I can speak from three years of personal experience as the lead for Project Maven, and now as the director of the JAIC.  While we moved fast in Project Maven, delivering the first A.I. capabilities to a combat theater within six months of standing up, there is no question whatsoever that both Maven and the JAIC would be much further along right now with A.I. fielding, had we had an enterprise cloud solution in place as originally scheduled. 

We have spent far too much time, over the past two years, striving to gain access to data for training our algorithms, updating fielded models, site by site, deriving ad-hoc solutions to bring real-world data back to allow dynamic retraining of fielded models, and cobbling together one-off, bespoke cloud solutions to meet mission requirements. 

For these reasons and more, both Maven and the JAIC plan to be among the very first projects to migrate to JEDI. 

Enterprise cloud provides a solution for the deployment of software and A.I., enabling instantaneous and continuous deployment of updates to a global enterprise.  A single instance of an A.I. application can be used by an entire enterprise, and updated for all at the same time.

An enterprise cloud allows A.I. cycle speeds that can be measured in updates and across an entire enterprise in hours as opposed to in months, six months or maybe even a year. 

The ability to run entire A.I. pipelines in the cloud will be critical for success.  From data acquisition to labeling, continuous training, test and evaluation, integration, deployment and even as far as human-machine teaming with continuous dynamic feedback. 

The top A.I. companies are, first and foremost, cloud, data-centric organizations with a continuous development culture where integrating, managing and analyzing data at scale is the lifeblood of the organization.  That's where we want to be and need to be with our enterprise cloud solution. 

The Joint Common Foundation, or JCF, as part of the JAIC that we're building, that is going to be an enterprise cloud environment.  It's a platform that will provide data, tools, environments and access to other certified platforms, to enable software and A.I. engineers to rapidly develop, evaluate, test and deploy A.I.-enhanced solutions to the warfighter. 

The JCF is designed to lower the barriers to entry, democratize access to data, eliminate duplicative efforts and increase value added.  That platform will reside on top of the JEDI infrastructure. 

JEDI will include cloud capabilities that are able to operate out of standalone portable hardware even in the absence of communications links.  It will resynch with the rest of the JEDI cloud as soon as communications are restored.  This hub-and-spoke model means that local military equipment that is connected to the JEDI cloud hardware could still operate and be used to execute missions in a degraded, disrupted or a denied environment, extending enterprise cloud, in other words, all the way out to the tactical edge. 

And finally, this is not merely an A.I. issue; It is also about joint all-domain war fighting, taking advantage of emerging technologies to develop new operating concepts for a kind of warfare that will look completely different than what we've experienced for the past 20 years.  

In this future high-end fight we envision a world of algorithmic warfare and autonomy where competitive advantage goes to the side that understands how to harness 5G, A.I., enterprise cloud and quantum, when quantum's available, into a viable operational model, all part of the department's transformation from a hardware -- hardware-centric to an all-domain digital force.  This digital more -- modernization is a war-fighting imperative that demands a palpable sense of urgency, and it's one that will be fueled by an enterprise cloud solution.

So with that said, I think we can turn it over for questions.

STAFF:  We can start from here. 

Q:  Okay. Thanks, Patrick Tucker from Defense One.

There's been a lot of public speculation lately, in part because of the president's Twitter feed, and also because things that you see on some news channels that -- people urging the president of the United States to personally intervene in this contract and cancel it.  So without getting into the politics of it, can you tell us what would happen in the event that JEDI were, by fiat or for some other reason, the program was canceled before being awarded, or asked to recompete?  Like, what would be the effects of those two, however unlikely, eventualities?

MR DEASY:  So all I can say is it is our job right now to educate the secretary as to why we need a cloud.  In doing so, we will make it very clear as to what it means to the warfighter, what it means to even, you could say, efficiencies across the Pentagon.  

In having that conversation, he will have a good understanding of what are the consequences, honestly, if we were to take a decision not to continue with JEDI.  But I can't speculate today, nor would it be appropriate for me to try to describe those consequences.

Q:  Okay.  How have -- can you give us a sense of how he would get a sense of the consequences?  (inaudible)

MR. DEASY:  So I'll -- I'll -- I'll -- so what we're going to do over the next several weeks is we are actually going to bring people in to these sessions.  So we're going to bring warfighters in.  We're going to bring the military services in.  We're going to bring combatant command participants in.  For example, U.S. Cyber Command, and we're going to bring technical and legal experts in, as well. 

So part of the education -- you know, I always believe that some of the best way you educate is to make it real world, and so we're going to bring people that can talk about the real-world implications and the importance of the cloud.  So that's how we're going to bring it to life for him.  

STAFF:  (inaudible)

Q:  Oh, yes, so in the -- at the beginning you mentioned your previous comments about August being the early -- the earliest.  Do you still -- is that -- is it still possible for August, or are you now going...

MR. DEASY:  No, as I said, you know, you have to remember how many months we were tied up on the federal claims side of that, which redirected a lot of energy and a lot of people's attention towards that.  Obviously, very pleased with the outcome of that.  But that work, necessary to do that, has caused us to have to, you know, spend now additional time to complete the evaluation process. 

One thing I've been consistent on from the beginning is, we've got to get this right.  So we are not going to rush to a decision.  We're going to spend whatever time that the evaluation team needs to spend, to make sure we are picking the best technical solution at the right price with the right delivery criteria that will meet the needs of the warfighter.  

And we already know, in looking at the timeline, what it's going to take to accomplish that, that we are going to go past the end of August.

Q:  Justin Doubleday, Inside Defense.  

You mentioned these sessions in terms -- you said weeks, weeks-long.  Do you have a schedule?  Can you say kind of like, you know, four weeks from now, you'll be done with this schedule of sessions?  Or what's the schedule? 

MR. DEASY:  So we have an outline, and I walked you through the outline.  What I believe will happen, just like any learnings happen, is that we will take him through, he will ask a number of questions, he will probably even ask a number of action items.  And we'll take out of that meeting, we will bring those back.  And this will be an iterative process. 

So right now, the reason I'm not going to try to predict an end date, it's all going to be to the level of depth we go into, to the number of follow-up questions and follow-up actions he'll give us.  And that's why it would not be appropriate to try to put a guesstimation on when that will complete. 

Q:  Okay.  And when you came into the department last year, I think you also did your own review of the JEDI cloud program.  And I'm wondering, what were you looking at?  What did you find during that review? 

MR. DEASY:  Yeah.  What I was looking for, was stepping back – saying okay, tell me how you went through the process of gathering requirements.  You know, how -- what -- what was the basis for this RFP.  And so a lot of time I spent earlier, I was understanding just that.  What was the warfighter requirement, what was the tactical edge requirement, how was it that programs like Maven was going to use it. 

Spent a substantial amount of time, understanding the security implications.  How was it, we were going to set up this.  What were the security requirements we were putting into the RFP.  Spent a lot of time, understanding how were we organized to actually do the RFP.  

How were we organized -- because you know, obviously, this was a new process for me, joining the Department of Defense.  How does the acquisition process work, what was the timeline for that, what was our selection criteria.  So those would be the things that I was looking at when I came on board. 

STAFF:  Do you have anything (inaudible)? 

Q:  Yes.  Emily Birnbaum with The Hill newspaper.  

First question would be, are -- do you think that you're going to bring in the top contenders, the companies AWS [Amazon Web Services] and Microsoft?  And the second question would be, could you offer some examples as to how this cloud solution is going to help on the battlefield, just...

MR. DEASY:  Sure. 

Q:  ... some like little illustrations. 

MR. DEASY:  I'll let Gen. Shanahan handle the second part of that.  To the first one, when you say, "Bring in"...

Q:  During this review process. 

MR. DEASY:  You mean to participate in the review?  No, absolutely not.  We -- we -- you should -- I should stress that we need to keep these processes separate.  Educating the secretary is one process, so he can weigh in and be informed in the direction he wants to go forward with this, which is separate from the sourcing selection process.  So those will be run separately. 

LT. GEN. SHANAHAN:  On the other side, so yesterday, I took -- I stay very closely connected to Maven, still.  And I got my -- an update yesterday.  And we were talking about Afghanistan.  And when I -- and I can't get into a lot of specifics in here, but the case they have is they're -- they're -- they're developing something called "smart system" in Project Maven, which is an A.I.-enabled ops-intel [operations-intelligence] fusion system along with another system that Special Operations Command is using on the operations side, how -- showing how you can bring ops and intel together.  

And the op tempo, the pace of operations in Afghanistan is so -- it's higher than it's been in an awful long time right now, and they were giving me very concrete examples of downrange using these capabilities.  But the -- the more they're using them, the more data they need, the more data they're actually pulling off the battlefield.  But there are specific instances of having to get all of this data -- there's no common data fabric.  Somebody has to actually clean that data, and if they want to use that to help then retrain the algorithm that's been fielded, they've got to do a bunch of steps to it.  They don't have time to do that.  It's -- it's site-by-site, and one of the task forces in a different location is bringing enormously important information and data, but there's no simple way to share that with another part of -- of Afghanistan, or another part of the task force.  An enterprise cloud solution helps address all of that.  That's one piece of it.

And to get even, let's say, more specific and to a place where JEDI is -- or enterprise cloud is fully fielded, that hub-and-spoke model -- let's say you have the central JEDI cloud, then you have, maybe, portable data centers that are downrange, a little bit -- you know, a bit bigger than sort of, you know, a breadbox, think CONEX [container express].  And then all the way down to a portable, truly, two-man portable cloud compute solution that anybody could stay connected to.  

The beauty of that is not only are you getting access to -- to all the benefits of the cloud down to the very edge of the battlefield; as you're collecting data, that data can then go back into that portable cloud, connect to the big CONEX all the way back to -- to the cloud, so everybody is benefiting from that.  And if you get disconnected, as is going to happen in combat, especially in a high-end fight, you still have what you had at the point it was disconnected.  When it suddenly comes back, you have all of this now being -- being connected across the entire enterprise.

So it's really -- it -- it really gives people access to more data, better data, faster data than one-off each site’s trying to come up with their own solutions.  That's a -- that's my quickest example that I can give you.  

MR. DEASY:  Yeah, I'll add one thing onto that.  It's interesting.  As I've gone out -- after joining the Department of Defense I had a chance to go out and meet with just about all the combatant commands, and they'll walk me through, obviously, their technical challenges they're faced with.  And I will point out, I'll say, "But we have a -- a number of clouds."  And I made a comment earlier at the beginning of this today.  We are not short on clouds in the Department of Defense, and that somehow has gotten lost in this narrative.  And they'll say, "Yeah, but we don't have an enterprise approach.  We have a bunch of siloed solutions we built.  We have lots of vendors we're using for cloud solutions, but we've never stepped back and created a holistic solution."  And that is -- that is causing challenges out in the field.

I've been invited, for example, some time over the next several weeks I will -- I will be going to Afghanistan, and the reason they want me to come out and spend some time with the forces is to actually see the problem of what he was -- what Gen. Shanahan was referring to by having siloed data, by having these clouds that have not been connected up in an enterprise manner -- to actually see how it's impacting the warfighter.

Q:  Good morning.  Thanks for doing this.  Amanda with CNBC.  

Sir, can you say if this is a break with bureaucratic process of awarding a contract?

MR. DEASY:  I'm not sure I follow.

Q:  So right now, with this particular contract, is it a break with the normal bureaucratic process of (inaudible)

MR. DEASY:  No, and the reason I say that is we have not stopped our selection process.  I think what JEDI is going through is common to other large programs.  You're going to have protests.  You're going to have -- you're going to have people that are going to weigh in.  We have a new secretary.  And as I said earlier, I would absolutely expect him, on a program of this scale, to want to completely understand that.  

It's not a break from the standpoint -- we got work to do, we're continuing with that.  We're not stopping anything.  While, at the same time, we are educating the new secretary on this. 

Q:  Is there anything else in your portfolio that you're aware of that Secretary Esper's going to get a deep...education session…


MR. DEASY:  Yeah, so I mentioned up-front, that we have this digital modernization program that consists of cloud as a foundation, A.I. as an enabler, C3 -- command, control, communications -- next generation of that, and finally, cyber. 

He is going to have a series of ongoing deep dives across all four of those portfolios.  So actually, all we're doing is we're accelerating his deep dive on the cloud.  

But we've got, coming up in the next month or two, a deep dive, for example, on the Joint Artificial Intelligence Center.  We just did one a number of weeks ago on command and control and communications.  So this is just one of a series of deep dives.  

Q:  Okay.

Q:  Aaron Gregg with The Washington Post.  Based on your conversations with Secretary Esper, is amending the RFP to go multi-cloud for JEDI, something that's on the table?  And do you have the workforce in place to implement that? 

MR. DEASY:  So, right now, we have not modified, changed anything in the work we are doing on JEDI.  We are still proceeding through the sourcing -- selection -- source selection process. 

All we are doing with Secretary Esper is educating him through that outline that I shared before.  At the end of this process, he will obviously be at a point and a position to weigh in with his thoughts and views on this.  We obviously will need to incorporate his thoughts and views on that at that time. 

But at this point, no, there is no changes to the direction.  The only thing we are doing is educating the new secretary while completing the evaluation program. 

STAFF:  (inaudible). 

Q:  I've got a couple questions.  How come you all didn't do this sooner?  I feel like this would've been a thing you could have done months ago that would have helped quell the rash of misinformation, disinformation out there.  So what precipitated this....

(CROSSTALK)

MR. DEASY:  Yeah, I will -- I will tell you, one of the things that we have to delicately balance is the source selection process, and what you should and should not say publicly when you are in the middle of a source selection process.  You're dealing with that. 

More importantly, when the protests started and we went through the whole series of the GAO [Government Accountability Office] review and then the Federal Claims Court review, during that time, we have to -- we have to be very careful about not getting ahead of or interfering with either of those types of processes. 

So it would have been inappropriate for us, in the middle of those, you know, GAO and Federal Claims Court processes, to have had the conversation we're having today. 

Q:  Sure.  And Mr. Shanahan, on -- one of the examples you gave when you were talking earlier, was about the deployment of algorithms across DOD's hundreds -- probably, or more -- networks.  Do you have other examples?  You talked a little bit about the tactical edge, but where JEDI would just basically crush whatever systems, processes you have now and make you guys more lethal? 

I think that's something that gets buried a little bit in some of the other stuff that comes up, but it's hugely important for explaining why we need this enterprise-wide cloud.  If there are other examples you can share, I'd be happy to hear it. 

LT. GEN. SHANAHAN:  Well, a couple of thoughts.  Day-to-day, my main interest, of course, is in A.I. because of the position I have, the position I held before.  And I've been down this path long enough and far enough to know, you cannot get to true enterprise A.I. solutions without an enterprise cloud.  That's what the biggest -- biggest AI companies are doing exactly that, just tend to do -- they also tend to be the biggest cloud companies and there's a lot of reasons for that.

But just the entire pipeline -- and -- and I just remember the days of -- of Colonel Cukor who led Project Maven sending a crew site by site by site to install that first algorithm, but even worse, having to go back site by site to then put updates into that.

And then how do we get the real world data back to make that algorithm better?  Because it always -- it always will change based on real world data, but that will all be made so much better, so much faster, so easier with an enterprise cloud solution.

It's not -- it's not going to completely eliminate the challenges but it gets a long way there.  But just in general to -- to your -- kind of the broader question here is any war fighting operation period having an enterprise cloud solution that has -- that allows interoperability -- right now, there are questions of cloud A is not interoperable with cloud B, what does it take to make that happen?  How do I get data from over here to data over here?

I think we're seeing real tangible examples of that playing out in Afghanistan right now and that's in a fight that we're largely calling counter -- counter-terrorism, counter-insurgency.  Imagine the speed of operations in a fight in the -- in the Pacific where you just do not have time to figure out how do I get my data, clean my data, move it from point A to point B?

If I am a warfighter, I want as much data as you could possibly give me, let me use my sort of algorithms to sort through it as fast as -- as -- at machine speeds, let the machines do that but the humans think, the cognitive piece of this.

I -- it's really hard for me to do that without -- without an enterprise cloud solution.  I mean everybody is exploring different sort of cloud solutions, but again, they're -- they're tend to be bespoke and what we're looking for is a joint enterprise solution.

The whole idea of joint all domain command and control which the joint staff J-6 has now been chartered to lead across functional team, that is going to rest on the backbone of a, you know, enterprise cloud solution.

Q:  And just -- just really quick, the -- the -- the point about adversaries ramping up their investment, we've seen China do this with investing $10 billion recently in their own A.I. efforts.  I wonder how important if you can -- qualify how this compares to other adversaries that we're dealing with now and how that ... 

(CROSSTALK)

LT. GEN. SHANAHAN:  We don't want to waste any more time moving forward because we know our potential adversaries are -- are doing it at -- at their own speed, whether it's Baidu, Alibaba, Tencent, SenseTime, they're all coming up with their own cloud solutions and -- and then don't get me wrong, I don't want to make them 1,000 feet tall, they're going to have their own cloud interoperability challenges.

But the level of investment and the number of people they're putting at the problem, they're moving at a very rapid pace and what I can't afford to do is slow down anymore.  We -- we -- we want to move equally fast and actually outpace them and they do -- they -- they're taking a very similar approach, they know the importance of an enterprise cloud solution to what they're doing with artificial intelligence.

And Russia's not -- not that far behind, a little bit different approach in how they do it, but in general, your point is well taken that we've got to -- we've got to keep our speed up.

MR. DEASY:  I want to bring this to light, you asked a question about how -- what this could really be as far as a game changer.  So just think about your own personal iPhone or Android device.  You get updates almost daily and that's because you're all connected to a cloud and those updates can be pushed down to you.

Think about something like Maven today and all of the various sites that we're now starting to put that out on.  And as we start to have new warfighter upgrades, new capability that help them in the fight, think about how we have to push that out today.

Because they sit on all different platforms, different clouds, not integrated.  Think how slow those upgrades go.  Think about if they were all on the enterprise cloud.  Now we can push out a new update to the warfighter and they can get it instantaneously, versus the approach we have to take today because we have so many disparate clouds.  

Q:  Good morning.  Ryan Tracy with the Wall Street Journal.  

What is the role of the commander-in-chief in deciding the future of this contract?  And also, why, for at least the first two years and the minimum $1 million, are you choosing a single awardee? 

MR. DEASY:  So we hoped today, just through all the examples we've talked about with the warfighter and the disconnected, siloed data non-integrated clouds, that that's very apparent that the enterprise cloud is trying to solve for the fact today we cannot give the warfighter the data they need on the battlefield in an integrated way, updated application capability in real-time.  We have to learn how to do that.  This is not going to be the only cloud we're going to have across the Department of Defense.  We are always going to have a number of clouds, but to do enterprise cloud at scale all the way out to the disconnected warfighter.

Remember what he – Gen. Shanahan said.  You've got to have cloud, and you're going to have a forward command, and then you're going to have the disconnected warfighter.  We have to learn how to start to integrate that, and then over time, we will bring more and more cloud providers into that.

Your other question about...

Q:  What is the role of the commander-in-chief in deciding the future of this contract's program?

MR. DEASY:  So there is a separate team that is responsible for the technical selection of the outcome of who we're going to use, and that provider.  That is separate.  His role is to fully understand why we need a cloud, why this approach, and that we have done this approach in the right, appropriate manner.  And then his role is to look at that and decide if any changes to that approach needs to take place.  That is his role here.

Q:  So he could have a role in deciding whether it goes forward or not.  

LT. GEN. SHANAHAN:  Does the commander-in-chief...

Q:  The commander-in-chief -- I'm sorry, not -- not the secretary.

MR. DEASY:  Oh, I'm sorry.  I'm thinking secretary.

Q:  I’m asking about the president.  Sorry.  What -- what's the role of the president...

MR. DEASY:  None.

Q:  ... in deciding the future of this contract?

MR. DEASY:  Not -- not in this particular contract.  I'm sorry.  I was thinking...

LT. GEN. SHANAHAN:  Secretary, yeah.

MR. DEASY:  Secretary.  You guys are probably all going, "What is he saying?" (Laughter.)

LT. GEN. SHANAHAN:  That would have (inaudible)

MR. DEASY:  No.

(LT. GEN. SHANAHAN ?):  Because -- yeah.

MR. DEASY:  Let me be perfectly clear. I was referring to Secretary Esper in all that.  As far as the commander-in-chief -- not involved in source selection process and not involved in all of these efforts that we are doing with Secretary Esper. 

Q:  Thank you.

Q:  Could he -- you know, he is the commander-in-chief.  I mean, could he say -- could he give an order to -- change it and -- you know, how does that work?

MR. DEASY:  I -- I would not be in a position to...

Q:  Okay.

MR. DEASY:  ... speculate on that. 

Q:  Jacqueline Feldscher with Politico.  Given the president's apparent interest in this, has there been any thought here at the Pentagon or any requests from the White House to do these sort of education sessions with someone at the White House, in addition to Secretary Esper?

MR. DEASY:  So you should know that for several months now, one of the roles that I played, as to leading this, is to go out and offer education sessions to a number of people across government.  So I continue when asked.  We'll field calls from Congress.  I will go over and meet staffers.  I will meet with whomever is asking to get more education on this.  

So for the last -- let's see, I've been here now, what?  Not quite a year and a half -- I have been giving ongoing education sessions – to people as they've asked for them.

Q:  And so if the -- the source selection review process finishes before Esper's education sessions are done, will you award the contract while these sessions with Esper are still ongoing?

MR. DEASY:  No, we will not.  Obviously, we -- we'd need to ensure that he is fully-educated.  Right now, that hypothetical question is highly doubtful we'll even find ourselves in that situation.

Q:  Thank you.

Q:  A couple Esper questions, and then for general.  As part of the transparency effort here under Esper, did he approved today's session and the fact sheet you put out?  Realizing he's been in Mongolia with horses but).

MR. DEASY:  No, did we -- did I personally sit down with him and discuss the nature of what we were going to go through?  No.  You know, did I say to him when we first met, "Sir, I believe we need to have further discussions with the press.  We need to clarify some of the myths that are going on out there, especially around the importance of this to the warfighter."?  I said, "Absolutely, we need to do that."  But did he go through this narrative today?  Absolutely did…did not.  

Q:  But he approved you going out to -- and conceptually doing this?

MR. DEASY:  He -- he is supportive of educating everybody on what it is we're trying to achieve here and the importance of the cloud.  

Q:  Now, overlaying all this is an enigma wrapped in a cloud -- the I.G. [inspector general] investigation.  Whatever they're doing, they have been -- they've been cagey on this.  It's ethical -- or violations, or whatever.  But to what extent is that a pacing factor?  Do you need some clarity end -- end report like, the Shanahan-Boeing report before you can...

MR. DEASY:  Right.

Q:  ... with a -- an award?

MR. DEASY:  So on that front, what I've said in -- consistently in -- in talks with Secretary Esper is the fact that before we take any final decision, we will obviously consult with the I.G. on this.  I mean, our role right now is to let the I.G. -- you know, they need to be independent.  They need to complete their process.  But if we get to the point where we do not have an I.G. report, before we would issue any final award, we will obviously have a conversation with the I.G.  We will consult with them and in having that consultation, decide if there's anything that they're sharing with us that would give us reason to pause before it can continue on with award.

Q:  Okay, that's fine.  General, on the -- can you real-world example?

LT. GEN. SHANAHAN:  Yup.

Q:  The world is fixated on Iran right now, and the Persian Gulf.  To what extent would a enterprise cloud help CENTCOM [U.S. Central Command] cobble together -- you know, NAV Fifth Fleet, Air Force's Central Command, ships in the region, commercial traffic -- I mean, how would it help share information?

LT. GEN. SHANAHAN:  Yeah.  Yeah, I would even make it more -- broader than that statement, is any combatant command in the world would be better connected with an enterprise cloud solution.  So rather than -- rather than -- we -- we tend, just the way we've been operating absent a -- an enterprise cloud -- so these individual clouds -- even when we have individual cloud solutions, right?  We're just getting into a world of that, versus stack by stack, hardware stacks somewhere downrange.  The idea of putting that together so different -- different fleets can communicate, different air forces, Air Force communicates with fleets, communicates with -- with divisions.  Yes, all made -- all made better by the ability to have an enterprise cloud solution.  

Q:  So...

LT. GEN. SHANAHAN:  And -- and I would say, you know, of course CENTCOM, but I wouldn't limit it to CENTCOM.  This is a global combatant command, warfighter requirement that Gen. B.J. Shwedo talked about, Joint Staff J-6...

Q:  Yeah.

LT. GEN. SHANAHAN:  ... when it was -- is when he was talking about the importance of JEDI to warfighters.  Same point that he made.  This is something that everybody will benefit from, and dare I say it?  That INDOPACOM [U.S. Indo-Pacific Command] maneuvering units can talk to CENTCOM maneuvering units through an enterprise cloud solution, which I say is not a trivial task, unfortunately.

Q:  Could I ask you, is Maven -- you mentioned, it's being moved -- is Maven being transported also to the CENTCOM region, the AOR [area of responsibility]? 

LT. GEN. SHANAHAN:  Yes. 

Q:  Iran region, the Iran

LT. GEN. SHANAHAN:  Well, Maven is in the entire Middle East region right now, in deployed locations. 

Q:  Part of this expansion...

LT. GEN. SHANAHAN:  Yes. 

Q:  ... Mr. Deasy? 

LT. GEN. SHANAHAN:  Yes.  

Q:  Interesting story in itself. 

LT. GEN. SHANAHAN:  Yes.  And so the very -- very briefly sort of, just to put a couple of these things together, is we look at, how do we get to a cloud solution with Maven.  It was an extraordinarily hard thing to do because we didn't have one.  We had to rely on USDI's [Under Secretary of Defense for Intelligence] intel authorities to take advantage of the cobbled-together solution using one of the existing cloud solutions, just for proof of concept. 

Now, I'm experiencing the same thing, building the Joint Common Foundation in the JAIC.  We've had to go out and take some really convoluted paths to get just the beginnings of some cloud so we can show, as Mr. Deasy saw when he visited us over at our location, to see what that common foundation is capable of doing. 

Without an enterprise cloud solution, this is -- this is far too complicated.  I can make that a lot easier by the cloud office, the Cloud Program Office, taking care of most of the administration pieces of this.  And we just go out and provision the cloud. 

Q:  Thank you.  That's clear. 

MR. DEASY:  And let's not lose the fact that when people think of cloud, they think of this storage environment, they think of this compute capability where all this data is stored.  I would say it's what you put on the top of a cloud that really matters.  

And the fact that we're going to be able to build applications in a more consistent way because we're going to now have a methodology and a road map by having an enterprise cloud to build differently.  The fact that you're going to build applications differently, which means that you can secure them in a very different way versus the multitude of ways that we build applications today on a disparate number of clouds.  And I don't want that to get lost in this narrative as well. 

LT. GEN. SHANAHAN:  And if I might, that answers part -- I'm going to answer part of what you were asking earlier, too.  I think this is a point that I didn't raise earlier.  If a unit is deploying somewhere, you know, let's just say into theater; it doesn't have to be on-ship, but let's just say Army, Marine unit, and they're taking a portable device, a compute device.  Each one of those today is built on a different operating system at a baseline.  So if you want to put some apps on top of that before you deploy, it takes an awfully long time.  

In the proposed hub-and-spoke model, those portable devices will have a commonality to them, but very easy to build apps on top of them and just deploy them as quickly as you need to, to get down-range.  

As Mr. Deasy was alluding to, as we build out a platform for the Joint Common Foundation, our whole intent -- I used the word or the phrase earlier -- is democratizing access to data.  There is no reason in the future why a Marine or a soldier in a forward location in a combat outpost in Afghanistan, can't do his or her own app as long as we've given them that common foundation, built on an enterprise cloud solution. 

Imagine that scenario, where I need something right now, the data's out there.  Today, they just can't access the data.  It's too hard to do, the data's dirty.  If we can get them to do that, they can actually start doing apps for wartime reasons in a much, much quicker timeline. 

STAFF:  Did any of those dial-ins join us? 

Q:  Yes, I had a follow-up.  Just on Jackie's -- to follow up on Jackie's question, have you actually briefed anyone senior at the White House?  Yeah, done educated -- I know you said you did Congress and ...

(CROSSTALK)

MR. DEASY:  I continue to brief a number of people across government. 

Q:  And do you have any sense of how long this education process for the secretary...

MR. DEASY:  I think that was the question Justin had asked earlier.  There is not an end timeline.  And one of the reasons it's difficult to predict, though we have an outline, just think about the way that you learn.  There's a series of follow-on questions.  I'm sure they'll give him -- he'll give us follow-on actions, things he'll want to understand deeper.  We'll come back to him. 

So I -- all I can say is, it will be an iterative cycle.  We have an outline on the approach, but I can't predict what the end date will be. 

Q:  I have a follow-up on my earlier question.  You'd spoken before about how having more than one awardee for JEDI would create unnecessary complexities, as you put it.  Does DOD, based on your current cloud workforce, have the people in place to do something like that?  If you were told by Esper tomorrow...

MR. DEASY:  Yeah. 

Q:  ... "We want to go multi-cloud," could you do it? 

MR. DEASY:  So understand, we will have a multitude of clouds.  What we're trying to  describe today is, if you think about everything just said – Gen. Shanahan said, think about the complexity of what it's going to take when we bring in the first cloud provider, to set up the unclassified, the classified, the Top Secret, CONUS [continental United States], OCONUS [outside continental United States], Forward-base, disconnected warfighter.  

Just pause for a second, think about the complexity of setting that up.  We've got to learn how to do that with somebody. It doesn't end there.  That doesn't mean we're going to end up with just a single cloud provider.  That's why I have stressed, the two-year break with a three-year extension, with a three-plus-two.  That was done by intentional design, that as we learn how to do this, then we can start to bring in additional cloud providers. 

To your specific question, we will source the appropriate talent we need for whatever the permutation, whatever is going to be the end state.  So we will make sure we're successful, in sourcing the right talent. 

Q:  Thank you. 

Q:  (Inaudible).  

So the way that this is structured intentionally, does this give DOD, in a way, a sense of pivoting if the first contract goes out for the million, say two years, and then two years later, someone else -- you could essentially recompete and have another...

(CROSSTALK)

MR. DEASY:  Yeah.  I mean -- I mean, you know, this is -- this is a standard procurement process.  And then at the end of the two years, we have decisions we could do.  We could obviously extend.  What we will do, I can guarantee you, is we will look at the state of the marketplace. 

I mean, one of the reasons that I chose the two, three, three, two is if you kind of look at where cloud technology is going, the natural maturity of the cloud marketplace is every two to three years, there is major breakthroughs.  

So it's very much by design, that we say we need these break points.  Because there could be new technology from a completely different player, that can help us in the enterprise fight that Gen. Shanahan talked about, that we're going to want to bring in.  

And we're going to use those natural break points to do just that.  Scan the marketplace, evaluate new technology offerings.  Based on whoever we have, decide if that's the best course to stay with, or to bring in multiple. 

You know, all you have to do is look at the evolution of technology.  And I think we can all safely say, the market's going to look very different, next two years, three years after that, three years after that, which is why we constructed it the way we did.

Q:  Okay.  So in terms of financing that, are you hoping for a more competitive price point, the next time you go back...

MR. DEASY:  So what I can't get into is, obviously, the source selection process.  Suffice to say that the way we are constructing this, we are going to get very, very competitive pricing.  And the way we've set this up, with the two plus three plus three, is it will always keep us at the forefront of the most competitive pricing. 

Q:  Thank you. 

Q:  Okay.  One of the narratives, or questions that people keep asking me about is this one of centralizing this entire thing with one provider, does that create a central target and thus a vulnerability going forward?  Can you address that idea?

MR. DEASY:  Yeah, so first of all, in answering that, there's a huge assumption there that everything we do across the Department of Defense can sit on this one cloud, that's simply not the case.  Think about the fact we already have a multitude of clouds. We already have a very large installed internal base of IT applications and systems.

One of the things that we spend a lot of time talking through is A, we've got to make it simple for people to get access to and to put new application capability up on to the enterprise cloud, but we have this ongoing tension of we've got to make it simple for people to use but we've got to make sure they're always using it in the most secure manner.

So there's a number of safeguards we have in our selection process through the RFP, and that we will be doing with the help of U.S. Cyber Command, NSA [National Security Agency] and DISA [Defense Information Systems Agency] that will constantly allow us to stress test our commercial environment to make sure it's secure.

LT. GEN. SHANAHAN:  And -- and I can tell you from my own -- getting my own education in the two years of -- of directing Project Maven, I got deep into the cybersecurity aspects of it, and then both on physical security to see what these clouds -- I mean their -- their entire business model is based on consumer confidence that their data is going to be protected.

So physical security is incredible and then virtual security in terms of terms I didn't even know existed two years ago -- sharding and striping and sending data out into -- and encrypting data at rest -- and then the idea of -- of 1,000 attack surfaces versus protecting a much more narrow -- I won't say it's only one attack surfaces -- there are advantages to cybersecurity having a -- a -- a limited cloud -- a number of cloud providers for that reason.

There's a -- nothing is ever invulnerable, nobody would ever claim that, but there are -- the -- the -- the money, the time, the people they place on -- on security across these enterprise cloud companies today, it's an enormous amount of resources they put against it.

And I think if Gen. Nakasone was here, he would echo that from his role as the Director of NSA and Commander of Cyber Command.  He knows -- knows a lot more about cybersecurity than I do and I think he would echo ... 

(CROSSTALK)

MR. DEASY:  Yes, on of the conversations Gen. Nakasone and I have all of the time is because of the multitude of clouds, these siloed, independent -- independently configured environments that have been established over a number of years that, where he said this is -- this enterprise cloud actually helps us to become more secure than the  environment you have with a multitude of clouds that are set up in a variety of different ways.

Q:  Can I ask you -- on the source selection process, I know you -- you can't get too much into it, but this building has had great success in the last few years of (inaudible), becoming protest-proof basically, not losing protests.

Can you give a feel for the size and scope of -- of this source selection team?  Is it a team of like 30 individual operations that will be ... 

MR. DEASY:  Yeah, that would be inappropriate for me to start to describe the team, the makeup of it, other than to say that they come from a variety of parts from the Department of Defense, they all come with the necessary expertise you would expect a team to have to be able to properly evaluate cloud providers.

Q:  You're not part of that team ... 

MR. DEASY:  No, I am not.

Q:  ... you're -- you're not the source selection authority ... 

MR. DEASY:  No, I am -- I am not.

Q:  And Gen. Shanahan is not?

LT. GEN. SHANAHAN:  No, I'm as far removed to that as I could possibly be.  Thank you.

STAFF:  All right, I think we have time for one more.

Q:  Well, can I have two really short ones?  The first is on IT, like workforce stuff.  I know you guys have a hard time hiring technical talent sometimes, a big issue for you guys.  When we talk to smart people from the valley and whatnot, that want to come work for the government, one -- their cloud native, they're used to working in those kinds of systems and they say it's a huge barrier.

Does this address that?  Would that address that so you might get a better pipeline of smart people coming in?  

And then secondly, when it’s operational, will you get a mandate down to the departments to ensure that they begin migrating and or would that cause some of the existing clouds to kind of go away to get under the same architecture?

MR. DEASY:  Yeah, so both of those -- the first one, one of the things I'm really excited about and I've seen this throughout my career is when you have a world of just about every type of technology platform -- and the Department of Defense has accumulated over the years just about every technology platform -- think about staffing, hiring, retaining for that variety.

Now imagine in a world where we now start to have a much more standardized environment with latest, greatest technology that is constantly being updated, that is highly attractive to young people, that's exactly the environment they want to come in with.
They want to come in and work with the latest technology, they want to be able to use those skills.  Those skills obviously we then have to work really hard to retain them, but that's what they want to work on.  And so this is a world that we want to create for a lot of reasons, one of which is just from a workforce standpoint, it's just a lot easier to manage and maintain those individuals.

Your part two question was?

Q:  When it’s operational issue some kind of mandate to ... 

(CROSSTALK)

MR. DEASY:  Oh, there's a government at large mandate and we call it a -- you know, cloud smart strategy across all of government of -- the philosophy that I have -- share with the Department of Defense is, you know, we have two worlds.

I use these terms, we have a green field and we have a brown field.  The brown field is all of the technology that we've created over a number of years.  Now, you should not hold that we're going to simply lift all of that technology and drop it into a cloud.  It simply does not make sense.

We are going to have to go through a very thoughtful process of what of our existing applications would make sense to lift and shift over?  We're going to look at such things as natural break points in the technology, where you are going to do upgrades on the infrastructure.

And so we're going to have a methodology of how you go through and select the appropriate existing applications to re-port to a cloud.  Then there's green field -- green field are the things we haven't created yet today.  These will be the new ideas, the new problems that we'll have to help solve for the warfighter tomorrow.

All the services right now are moving towards a cloud first mindset that says when you're going to start something new, the bias will be to build it onto a cloud first.  It could be JEDI or it could be a unique environment based on the requirements of the warfighter, but yes, our bias will be in the future to move to a cloud-first mindset.

STAFF:  All right, thank you gentlemen, is there anything you wanted to leave the group with?

MR. DEASY:  No, just thank you for your time and -- and hopefully this helped to dispel some of the narrative out there.

LT. GEN. SHANAHAN:  Thank you.

MR. DEASY:  Thank you.

-END-