Header Image

NDIA Conference

General James Cartwright
Washington, D.C.
3/6/2007

Topic: Collaboration at USSTRATCOM

General Cartwright: I appreciate the opportunity to come and talk this morning. This network-enabled environment that we're in is really important to me and to the command. Given the number of missions we have and the global span of activity it really is not doable without moving in this direction.

The world we live in is certainly the information age and we're all well aware of that, but probably on the business side more so than maybe on the military side. You are not going to gain competitive edge unless you know how to get yourself distributed, lean, and flat, and [inaudible] agreements, way of looking at globalization and all of those things. But the reality here is that the decision cycles of today are a lot faster. Listening to some of the questions that were wrapping up there towards the end about processes and what not, there might be just a slight mismatch between our budgeting process in DoD and Moore's Law. [Laughter]. The one thing you can be assured of is that you can label your project, whether it's the F-22 or the next ship, legacy before you get to IOC. That's really not good.

It's not that we don't need exquisite systems, that's not the issue. It's that we need decision cycles, we need adaptability, we need to be integrated with our infrastructure in a way that lets us stay ahead of the adversary in decision time wise, in adaptability, et cetera. We can't do that with iron mountains or inventory. That's not the way to manage risk any more. We've got to do it with an adaptable, flexible way of doing business and that's only going to come through network enabled activities that are matched up with the infrastructure and the people that can work them. Absent any part of that triangle and you don't have a way of doing business.

Having said that, it's been slow at best in DoD. As hard as you can push, it's really not the technology most of the time that we deal with, it's the cultures. So for me, working inside of STRATCOM, trying to get people to move to a decision cycle that's actually relevant instead of perfect information that's irrelevant, is a big of a challenge. So I generally will set up the Joint Staff and the Office of Secretary of Defense as the enemy and say if you can't out-think them, let's just give up. [Laughter].

Sometimes it gets their attention.

I want to step through kind of how we're organized at STRATCOM, why we have organized the way we have, and it's a bit different than the conventional model, what the objectives are, how we're doing, a little bit on that activity, and then step into a little bit of how we're trying to get those adaptable, flexible capabilities and what we're doing from the standpoint of management of data and some of the other things to move ourselves forward.

The organizational construct, you heard all of the missions. STRATCOM started, they pulled together the Navy and the Air Force, what used to be Strategic Air Command in '92, put those together so we had what was called the old triad which was the bombers, the ICBMs, and the submarines. When we got to 2002 we brought space. We deactivated Space Command and brought that in under STRATCOM. In 2003 we had a fire sale and we picked up missile defense and ISR, global strike. In 2005 we picked up combating weapons of mass destruction. I'm hoping in 2008 we'll get the world hunger piece. [Laughter].

But we brought all these missions together because they were the types of missions that were cross-cutting. They were enablers for most other activities. They were key enablers for regional commands. So the idea was to start to bring them together. If I was in business I would say that STRATCOM really deals in arbitrage. We work the seams. We try to create the enablers and we don't try to close the seams, we try to exploit them. I tend to like that side of the equation rather than the defensive side of the equation. So I'm looking for seams and I'm looking for opportunity to break out in those seams and create advantage in an asymmetric way for the country and using these mission areas to try to understand how to operate globally to enable the flexibility that you really need on the local end which is the regional combatant commanders and the adaptability that they need. So I'm looking at scale, they're looking at flexibility and adaptability. They're the experts.

When we put the group together, STRATCOM in 2002-2003 timeframe was about 4,000 man years worth of work and a headquarters. Clearly an adaptable, flexible organization – Not. [Laughter].

What we did was set up a set of first order principles that we thought would take us in the right direction. The first is hard to scale, but you've got to be joint. But joint on the way to combined because we don't fight without allies any more. The more we practice not fighting with our allies the more we disadvantage ourselves. So why not go find that seam and exploit it? So the idea here is joint on the way to combined.

The second was to get ourselves in a position to be distributed. Not to be centrally located in Omaha, but to get out and amongst. If our customer base, so to speak, is out there and global then get yourself distributed. Find the centers of excellence and take advantage of them, exploit them in a way that allows the maximum amount of leverage that we can get.

The idea was take the 4,000, trim it down to 1,000, and there's a little bit of culture shock, and then distribute to these mission areas about 200 people each and try to take advantage of centers of excellence that already exist rather than trying to recreate them.

An example is for intelligence, surveillance and reconnaissance. My commander is the Director of the Defense Intelligence Agency. To make it joint, his deputy is an Air Force two star, a recovering F-16 fighter pilot who knows nothing about ISR but he's a planner. He's worked in Air Operation Centers. I don't have 200 intelligence people in that organization, I have 200 military planners that are inside the Defense Intelligence Agency to leverage that entire enterprise for the warfighter. So I'm trying to get the economy of scale of investing 200 to get about 20,000-30,000. That's what we're looking for.

The same goes across the mission areas. Set them up that way, leverage them, don't go in there with the same expertise. Go in with what you're trying to get on the output side of the equation, not the input side of the equation. So each of our mission areas are organized that way. They're also distributed.

These commanders have more authority than I'll ever have as a combatant commander. They have acquisition authority, they have requirements authority, they have resource authorities, all things that I have never gotten. So get that stuff out, get it distributed, get it down to where people can actually use it, use it in timelines of decision lines, that makes sense, and we'll support the warfighter. Get as much leverage out of the organization as you possibly can.

That's been tough, both on the reconstitution of the command functions and organizing the command functions, on the pushing of people out into distributed organizations. We in the military in particular, but it really can be generalized out to almost anybody, if you don't own it you can't count on it. It's the mindset. The reality is you cannot afford to do business that way. Business has learned that, DoD has not. We've got to. We've got to establish the trust. We've got to establish the fact that because it's not under your roof, because you can't reach out and touch it every day, does not mean you can't count on it, does not mean that it can't be responsive, does not mean that you can't get economies of scale and be able to do business under duress. All of those things are myths that you have to break down. And they're touched. Especially when you throw in service cultures and all the other pieces. I'll give you some examples of where some of these challenges have emerged as we've worked our way down through this process.

But organizationally, that's the way we're set up. Leverage off of existing centers of excellence. So ISR is the Defense Intelligence Agency. For our net warfare and for our information operations we use the National Security Agency. It's going to be the repository. You've got the center of excellence for high power computing pathology, mathematicians. It's all there, it's going to be there, that's the center of this kind of work for offense and exploitation in the cycle world anyway.

The defensive side is at the Defense Information Services Agency, DISA. Charlie Croom, I heard his name mentioned earlier. He does operations and defense for me. We're still trying to push this together. We've got to have integrated offense and defense. Anything short of that, it's not going to work. It comes slowly. It's not the way we're organized, it's not the way we're organized in the department trying to get this to come together. It's been a bit of a challenge. Right now I rely heavily on personality, but eventually that's got to be overcome and we've got to get to a cohesive way of approaching this mission area that is a little more effective than trust me, we'll get along together. But that's coming together. That's that piece of it.

On the integrated missile defense side of the equation, we reached out to the Army and Space and Missile Defense Command, a three star Army general that runs that for us. Most people look at that as the ground-based interceptor and the crest and the cost and all of the things people have been debating about that. You're missing the point of missile defense if that's what you're looking at. The reality here is that what missile defense is really doing is starting to integrate multiple sensors, multiple command and control systems, multiple weapons in ways that we've never been able to net them together in the past. It's still not really on the plane that I would like to have it on in that it tends to be almost paralyzed match-ups. In other words, this command and control, this weapon, this sensor. We've got to start to spread that out and I'll talk a little bit about what we're doing to move in that direction.

Global strike. Took 8th Air Force. Took the three star at 8th Air Force, which was obviously Air Force; put under him a Navy deputy, an admiral; and an Army chief of staff. Pulled that together. They do the integration between all of the elements, all of the various mission areas. They give us the global perspective. They net this all together for us.

Weapons of mass destruction, the pillars are counter-proliferation, non-proliferation, consequence management. That is the Defense Threat Reduction Agency. A little bit of a challenge here. A civilian can't possibly be in charge of military organizations so we had to call him a director because I have a civilian that runs that. He does just fine commanding as a director. [Laughter]. But he's there. He's got a Navy deputy. We're pulling that mission area together.

All of them are basically organized that way. In order to make this work you've got to get your distributed activity, you've got to get collaboration going, and so that's the first pillar that I'll talk about is collaboration.

We all do it. We call each other, we have meetings, we sit around, we debate, we figure out where we want to go, think about it, come up with courses of action. Totally irrelevant to the timelines that we're dealing with. I'll give you a sense of the timelines that I lay out for my commanders and then we try to build to.

Anything that comes off the face of the earth you have about 100 seconds to type it, figure out what it is, and act. I can't even get a phone call through that fast. But the national system is set up to have a phone conference about that. You try to do that in the middle of the night. You try to do that in the middle of the day, get people out of meetings. It's not possible.

In that 100 seconds, what do we do when we get people on the phone? We describe what's going on so we spend most of the time in discovery rather than in options and activity and execution. We can't do business that way. That's the simple one.

Virus launched out of Baghdad towards the United States, out to geosynchronous orbit, 23,000 miles out, and back down to Seattle for the attack – 300 milliseconds. I can't afford to do business the way we're doing business so we have to build the organizational construct to work in these timelines. We have to change the cultural approach to doing business, to having people get involved and discuss and review and then decide and then execute to intervention by exception – machine to machine, intervention by exception. Build the business rules. Missile defense won't work without it. Space does not work without that. We're at a huge disadvantage if we think of it otherwise.

But what happens to all the people who think they have a vote? They're disenfranchised. Business has discovered this. What do you do with middle management in those kind of timelines? It's a huge problem. I don't know that we've solved it. We can say that we're flattening the organization out but there's an awful lot that has to be done before we try to get to those timelines. But that's the information activity.

So collaboration for us, when I got to the command, in my previous commands generally what we would find is that in the morning the commander gets his situation awareness, information intelligence update. Thank God nothing happens for at least 24 more hours because that's the next time you're going to get any update. And bless those guys that come in at 2:00 and 3:00 o'clock in the morning to build the PowerPoint brief for you. It's absurd. That's not the cycle that we're operating on.

So trying to bring the tools to bear that allow you to have real time, on demand knowledge more like My Yahoo, Google, et cetera, than PowerPoint is where we've got to be.

We've introduced a lot of tools. We've partnered up with Joint Forces Command. People understand dynamic templates that can self-populate out of things like TurboTax, et cetera. They can be used in these environments. And chatrooms and blogs, even though when I started blogging it just about drove the command nuts, but it has great value, and VTCs, et cetera. All of those are tools. I'm not trying to exclude any one of them but they've got to be put in the right context for the right functionality. What tends to happen is they've become novelties rather than being turned into weapons which is what they are for me.

So chatroooms are great when you're in execution because everybody's sitting at their desk and they're executing, but they're horrible if you're not in execution. If you're just in situation awareness or planning, they just don't do a lot of good for you.

Blogs tend to be very good in those environments. You come to your desk, you work it when you're available; the enterprise works it when its available and you can move and exchange information on a global scale in ways that we've not been able to do in the past.

The problem with the blogs, and I see some familiar faces out there, is that people treat them like the chain of command. I can't put something into the information environment until I staff it which is just 180 out of where you want to be. We went through this and it's kind of fun to look back on it. The first six months we had a few people blogging and it was kind of the same people all the time on events. So you'd see something like a period of interest or a period of activity gets declared and people start to focus on a certain place on the face of the earth and information starts to flow very quickly, then it slows down. You go what the heck's going on? Well, I can't give you this until I clear it with my boss. Let me help you out here. Either you give it to me now or I fire you and your boss. [Laughter]. We had a breakthrough. [Laughter]. Six months later we retrenched. It's culture. It is. We had to break through again.

We went through what's called the tethered goat phase. Lance Corporal Cartwright. The commander made the decision and then he asked Lance Corporal Cartwright to blog it so that it looked like everybody was participating. We kind of fixed that. We worked our way through that.

But the idea here is, it's not your rank, it's not what you wear on your collar, it's your contribution. That's the measure of merit. It's not the chain of command. It's the chain of information. They're not the same. If they're the same, I can't win, and that's tough.

We worked this internal to the command, and it's really not about the technology. I can tell you about the applications. That's not worth the time. It's about the culture and teaching people and advantaging people and showing them how they can advantage themselves by working in these kinds of environments and applying the right tool for the problem. It worked inside the command for a while and that was fine. Then it started to spider web our way out to our components, went through the same learning processes. It tended to work on six month cycles, but we worked our way through that. Now we're spiraling out to the services and to other combatant commands. Over the last year the department moved to something that it had not done in the past based on this kind of press, which was have multiple combatant commands involved in the same exercises. Something we've not done before. Boy, you want to talk about a whole bunch of Type A's trying to run the same circus. Big challenge, but it's starting to work.

This year our objective is to get to the allies. We're in the first stage of real time, machine to machine exchange of information. And oh my God, we're going to give away all the secrets to the people who fight and die right next to us. Give me a break. Let's get this back on a level playing field.

We'll work that through this summer. It's not without technical challenge, but that's the small stuff. When I talk to my counterpart in the UK and try to work through this with them, it's interesting to listen to him tell me the same things that we've gone through, listen to him tell me about well, the Defense Department side of the equation, we're doing pretty well at the headquarters but JCHQ won't tell us any of their secrets. Sound familiar?

The collaboration activity inside of our interagency side has really started to pick up this year also, based on value. We have a very close partnership, and I would not have thought it would have gone this way, but the first people to step up to the table and say we've got to exchange information real time with you was Justice, which really kind of surprised me. They have very different rules about their data. Very different from what DoD has. But the reality of the decision cycles is about the same. And the regret factors of not deciding in time are really very similar. So they've been a pretty close partner with us right down to the standards level over the past two years.

Homeland is starting to come on board, the DNI is starting to come on board now in a big way, mostly out of fear of being left out. But the reality is that when they do come on board they either find value of they don't.

In the command we work on the Disney principle. You either have a line at the door or you shoot it and kill it and move on. If you can't create a line at the door, then get out of the business.

So as these people come on they either find value or they don't. We move pretty quickly on this. But this summer will be a big activity for us in trying to move interagency and outside to our allies with our information distribution activities. To me that's critical. That starts to realize the goals that we have and how we do warfighting.

The next piece which really underpins almost everything that we do is the data strategy. I'll be the first to tell you that we're rookies, but we're trying to understand how to move forward.

What we have attempted to do was to start at a very rudimentary level and figure out how to connect dots. The first thing we learned was that almost everything we have in the department that moves information is biased towards a known user. We know who's going to use that data so we template it as close to the source as possible and then move the template to the user. You can be inside a cockpit and you've got a radar and that radar's function is to drop bombs or shoot missiles. It templates the information, displays it to the pilot, and everything else is thrown overboard. If you're not a missile or a bomb the information is relatively useless.

The problem is that when you're working in the seams, when you don't know who your adversary is, when you step across the line of departure and shit happens – sorry. [Laughter]. You don't know how he got on your left flank and you're going to have to change and you sure as hell can't wait five years. You've got to make connections in the data in real time, or at least in days and weeks.

This past summer the North Koreans were nice enough to have a missile exercise for us. The first thing we learned when we put the whole system on line of missile defense was that we had no way to coordinate scheduled maintenance. We were on line for a little over 90 days. Just take large terrestrial radars down and calibrate them and do maintenance, to take the Aegis radars down, the forward based – all these sensors have to have maintenance. If you put too many of them down at the same time your kill chain's gone. Not to mention that you're crossing nine time zones and four different combatant commanders, all of whom think they own those sensors. Not to mention the captain of the Aegis who's really distraught when you're managing his energy distribution and his radar activity. It took us four days to connect the information together in a simple program that portrayed it and allowed us to schedule the maintenance without disturbing or losing a kill chain at any time.

If I'd had to do that against proprietary systems, which that's not a four-letter word, but if I had to go back and beg a company a third or fourth time to repopulate the same information they've done for someone else, then it would take forever. We've got to find ways to move quicker. That was an example of how you could, but that's got to be the way it's done normally, not by exception.

So the data strategies are critical. What we have done with the combatant commanders, we being STRATCOM and Joint Forces Command, is sit down now and say in partnership with the NII, what are the standards to register your data, to make it discoverable, accessible, and useable? What do you have to do? We've taken that, we've gotten checklists now put together. We've actually given them to the communist testers – that's OT&E. [Laughter]. And they've looked at it and said yeah, we can actually test to this. We've also given it to those who own the legacy systems and said shoot, it wouldn't take much to actually adapt our systems to be able to do this. Okay?

This is where kind of the crazy Marine comes in, so my edict is on 90 day centers you will start turning databases in the Department of Defense, the Director of National Intelligence side of the equation, Justice, and eventually Homeland when they come on board. So every 90 days we're going to turn X number of databases that we have prioritized based on output functionality of what we need and what we can do with it. The first 90 day center started in January and every 90 days we're going to do that.

Now people are oh my God, it's going to cost a fortune, it's going to take forever, you're going to be invasive. When you get into it, it really doesn't cost much money at all. You take legacy systems, turn them around, you're able to actually incorporate the data, use the data in ways that we've never thought about using it, connecting the dots between different phenomenologies, crossing what sometimes are called stovepipes. I refer to them as vertical cylinders of excellence so I'm not pejorative. [Laughter]. But crossing horizontally between those where it makes sense. Connecting any shooter, any sensor, any command and control necessary to have the output that you need. But the idea is get the data organized and discoverable in a way that you can actually advantage yourself. Everybody goes oh, my God, the security problems, the – The reality is you're going to need all new legislation. It's not the case. You just have to be innovative in how you manage the data.

So we've got now these sheets, we call them purple sheets, that we are holding people accountable for incorporating and to which we will test whether or not you meet these standards. And it's actually really starting to take hold.

The next layer beyond that is to look at functionality and decide on the output side of the equation what do you want to be able to do? For STRATCOM, what seams do you really want to get into and leverage? For that we've taken and kind of stolen and bastardized a little bit communities of interest. But they are focused on an output function, not on an input function. I'm not interested that we've got all the maritime stuff in the community of interest. Who cares? I can't do anything with that on the input side. On the output side, do you want to be able to do something like [inaudible] strike. Sure. Do I care whether it's maritime, space, air, ground? No. So strike. Strike is a community of interest.

What we're trying to do in these communities of interest is tie this data together and build a known set of quality service activities that match up with the output function. So if it's a fire control solution the quality of the data, the timeliness, legacy, all of those things have to match up to the need of the customer and get them organized.

Well, we've been going at this for about seven months now. At the six month part they came to me and said one of the key tasks we've discovered is coming up with a reasonably common vocabulary. They really felt good, because across all of the services in six months they had eight words they agreed on. [Laughter]. You can imagine the baseball bat came back out. But it is hard. It's cultural. I'm not as a command trying to tear down service culture. I don't care whether you call it a wall or a bulkhead. It really doesn't matter to me. I don't care if your definition of a sortie is a takeoff and landing versus leaving port. But it really does make a difference when you put this together in planning and execution tools, and when you step over into the intelligence side of the equation or over into the justice side of the equation. How do you make the wrappers that connect this vocabulary so that it's useable in a quality of service that probably is measured in milliseconds of legacy. That's the piece that's critical.

So we don't go after this just for the fun of going after it. We go after it where there's leverage to give us advantage and we need your help in doing that. But the more process there is, the more databases, the more people have to review and mother may I's, that's what we're trying to get rid of in these communities of interest. That's what we're after is functional output capability, and it's critical. It's working.

Each of our mission areas have now developed their communities of interest. Each of them are working their way through it. When we start, we start with something called an engagement sequence group but it is the marrying up of a command and control system with a sensor, with a weapon.

The weapon could be something that goes bang. It could be something that prints a newspaper article. It could be something that just influences somebody. It doesn't have to go bang. But you've got to connect the dots. That's the rudimentary approach. Give me some capability in that area.

The real direction that we're heading in is called a kill web. That's when you decide that what you want to be able to do is conduct strike. Here are all of the service and other types of command and control systems, here are all of the sensors, here are all of the weapons. How on the input side do I find a common thread to go through that gives me the greatest advantage, so in missile defense how do you take space-based, non-imaging IR, connect it with terrestrial radar, connect it with a kill vehicle for end game? How do you connect all of those together? Different phenomenologies, different owners, all of those things. Quality of service has to be very specific. Put that to a command and control system and a weapon. Then after you've done that for one engagement sequence group, now how do I decide what's the next leg in this activity? Do I go after a new sensor? Do I go after a new weapon? Which one advantages you the most? Look at that and say gee, if I add this sensor in, let's say Aegis, then I could actually step across from long range ballistic missiles to short and medium range ballistic missiles and I'd be tied into the original chain. I could go back and forth. So one sensor could work across three different weapons or three different threats. So I zero in on that.

But it's a taxonomy that you put together to start to understand on the input side where's the greatest value for your investment? On the output side, where do you get the most adaptability, flexibility, highest probability of success? And in real time, when you're dealing in a global enterprise, how do you build ad hoc relationships in real time dynamically when somebody drops off-line because they have a maintenance problem or a schedule problem or they get shot between the eyes? And not lose opportunity, not lose advantage. That's where we've got to be.

So these kill webs are absolutely essential, but they go cross-service, they go cross-medium, they bust about every vertical center of excellence that there is out there. By design. But you're looking for advantage. You're looking for leverage.

For me, if you can't – I'm telling the Missile Defense Agency and others – if you can't give me a factor of five improvement, I'm not interested in your upgrade. It's just not worth my time. That bothers [inaudible], but generally you will get somebody to line up at your door if you can do that. That's what we push for, connecting these dots. It's critical.

I think the last irritation that I'll put toward you, and hopefully by now I've at least touched everybody's pet rocks here. But the last one I'll hit on is distribution and moving the information around.

The budgeting system in the department and the requirements system and the acquisition system really love exquisite. Give me something I can turn into a works project over years any day versus something I can just go buy off the shelf and incorporate now. We need a certain amount of that.

On the distribution side of this equation if, just if, if we were to go out and buy a space-based system that could solve every information distribution problem through-put, bandwidth, et cetera, known to man, we could spend most of every penny that we have to do it and take at least 10 to 15 years to field it, if we ever do. Would you take that investment or would you take that same money and get the 50 percent solution on the ground, 50 percent solution in air, and a 50 percent solution in space? When I look at that equation, boy would I like to go after the adversary that put all of its money in the first option. One ASAT would put you out of business if it's space. If it's air, I can do the same thing to you. We've got to have resilience in these activities. I tend to irritate my counterparts about this discussion, but I don't mind only having 50 percent throughput capability or bandwidth capability as long as I've got a capability that I can count on. I can't count on one answer to fix everything. Never will be able to. Not willing to make those kinds of investments.

It doesn't mean that we've got a huge problem in distribution, but I can't go that direction. That would be point number one.

Point number two is that in distributing this information we have not yet figured out a good taxonomy about who does what. One group takes the ten gigabyte file and sends it to a headquarters. The headquarters turns around and says I've got to have the ability to get this down to this shooter. Yeah, what are you smoking? That's not legal to grow in your backyard. You can't do that.

The commercial that I like, and I'm not supposed to use commercial names so I won't say Nextel, but when you watch that commercial and you see the guy managing the dots. Don't play with my dots. Dots are good people. Don't mess with my dots. Sniper Cartwright just needs a dot. I just need to know where the target is. It can be text, it can be geo-rectified on a PDA, I don't care. I don't need the ten gigabyte file. The services are good about understanding what Sniper Cartwright needs versus Pilot Cartwright, versus Sailor Cartwright, et cetera. They can do that for you. And the commercial sector, the technology available will drive that, innovation will drive that. If you try to go the other way you'll cut innovation off because you'll over-standardize the thing. You'll eliminate any opportunity to make advances and you'll template the information way too soon for the number of users you're going t have out there anyway.

So we've got to figure out how to get at this issue and start to mix thick and thin client and let the services work their way through this type of activity to figure out how to get the meaningful information out to the edge where it's needed. We don't do this well. We expect someone else to do it for us. We offload it to the wrong people when it ought to get out to the guy who says well really, all I need to know is X. Let me give you a system for X. It doesn't require a Humvee to power it and 50 vans of a DCGS or JTAGS. Some of this stuff is crazy.

The good news is, in my view, the services are starting to come to that realization. They really are. We're starting to partner up. We've got very rich and dynamic partnerships now with each of the services on moving their information and starting to break down some of these walls and getting it out there and getting them the information that actually advantages them, not what somebody has decided they should have. That's hard. We're still kind of, let's just say the military is still in the MOPL age and we are the producers and we're telling the customers what they need to know. We'll get this turned around eventually but it's going to take some time.

I'm hoping that I stepped on at least a few toes and got you thinking a little bit. It's really a pretty neat job. It's a huge sandbox, lots of opportunity, a chance to really get out there and move and push on things for a few years to try to get things to move in a direction that will advantage every soldier, sailor, airman and marine that's out there. That's our intent. I do work in the seams. We provide global services. Very seldom are we a unilateral actor out there. It's not my intent. Our intent is to provide services out there that will advantage the regional guys and allow them to adapt and change at a pace that they need to to stay inside the enemy's decision cycle.

I'll turn it over to Q&A. I think the ROE is anything you want to ask is fair game.

[Applause].

Question: You bring a lot of enthusiasm and a lot of change, and with your personal drive you are shaping this how people are operating, changing concepts of operations. But how do we sustain that when you move on? What is the governance process that sustains that?

I'll give an example. I was the Air Force Program Manager for the DIV several years ago when it started. [Inaudible] just now that we're thinking that that technology is diverting from a common interoperable baseline. So what's the government's process to sustain the change that we're trying to implement?

General Cartwright: A lot of times as you look back, change can be managed and manifested in some sort of a leader that comes into an organization that says we're going, this is the direction we're going, get behind me, and follow. In the military because of the, let's just say about every three years we tend to pack up and move on to the next organization or the next job, a lot of that activity turns into an activity trap and it doesn't get sustained to the next gen.

I went out to business and I went out to academia to ask that question and see what would they recommend when I was coming into the job, what would they recommend to try to get at this issue of will it be sustained. That's where this issue demand a factor of five improvement on anything you're going to change came from because that is so compelling it goes beyond the individual. If you can start to realize those types of metrics, then they don't care about you any more because they can go out and as long as they pass that gate the whatever it is, the activity will sustain itself.

I'm sure, absolutely positive, that the names of organizations, the names of things are going to change. But there is an inherent value in what we're doing, number one.

Number two, build the organization so that it can accept and embrace change, and that's a big challenge in the military, but if it doesn't then the retrenchment activity will go on for sure. But if the change is so compelling that people go gee, not only does this make sense but I could be a hero at the end of this, then they find a way to move forward. And absent being able to show them that, they will retrench back. So that's kind of –

I don't know whether that will work or not, but in the organization it seems to be having an affect.

On the governance piece, you have to routinize things. It's our comfort zone. It's how we understand how to participate and how your work force understands how to participate. Right, wrong or indifferent you'll grade the homework eventually, but my approach as an individual and a commander has been take the first 18 months and no worry about process, get the vision out there and get people moving and find the quick wins. Then start to build the process in behind it. Don't let anybody allow the cement to dry. Accept partial solutions. Accept the fact that you're going to irritate the Joint Staff and OSD. But move the organization and move the culture and then fill in behind with process but don't let it set up on you.

We'll see. I really enjoy going out and benchmarking against industry because unlike us, they either do it right or they perish. They generally have pretty good insights for you. It's not that DoD ought to be run like a business, but there are business principles that ought to apply to DoD and we're slow to adapt it.

That's a little bit on how I think about it. It will be interesting to see how this plays out, but my sense is the value is there. It will be tweaked, it will be played with, and it ought to be, and it ought to be set up in order to be able to do that as we move forward.

Question: What [inaudible]? [Inaudible] talked about some data out there that [inaudible]. Some you know that exists but they won't let you have it. But I'm asking about a third level where they just haven't done the homework. My question is, STRATCOM [inaudible] has not succeeded in motivating the [inaudible] foundational, heavy lifting to do network topology mapping including [inaudible] before you go to war. Who's [inaudible]?

General Cartwright: Yes and no. If you want to do that for the whole world, we don't have the scale yet to do it. If you decided that there are ideal or high value nodes that you'd like to have that kind of definition on, then we're able to do it. What has not been applied yet to cyber that we have for our general purpose forces and our strategic forces, so to speak, is both what's called an availability and a sizing construct. How much of it do you need and how much of it when are the two key questions.

So in our strategy today it's 1-4-2-1. That tells me that I've got to defend the homeland, I've got to remain aware in four critical regions, I've got to be able to conduct two simultaneous short fights, and I've got to be able to do a regime change. I size my military and capability to do that. That has not yet been done for cyber. We're trying to understand that, and as the services come on board, the military services, and start to integrate through our components with both the defense intelligence side of the equation and with NSA and CIA in particular to start to understand how much of this capability do we want them to have, how pervasive do we want to be, we're starting to understand the scale activity, now how do you train people – recruit, train, retain, all of those things – to be able to do this? How many of them should be empowered under Justice titles, Homeland titles, 50 in Intelligence, 10 in DoD. Because all of those, essentially what we're finding is because you're living in a millisecond world, when you have a firing line what you have are each of those titles represented on the firing line so that you have the right authority matched up with the problem at the time of transaction. It's not a bartering session. You can't go debate this. It's over before you can do something like that.

What we're trying to understand now is how big should the schools be, how much do we want to have of this, where does it give us the greatest value? I'm just talking about the mapping out of networks and understanding what's out there. It's more than we have the capacity for today, for sure. How much are you willing to trade of something else to have this? That's really the argument that's going on inside of the services today. It's okay to stand up and go I'm the Cyber Service. Show me your checkbook. Show me the people you're willing to train and take off of an airplane, a ship or a tank. Then I'll understand better how much you really value this domain. But let this activity demonstrate to you how valuable it is in comparison to an airplane, a tank, or a ship. That's the debate, the discovery activity that's going on today in the services. I will tell you the intelligence community is a little further along than that. They really are. They've thought through this. They don't have as many people as they want, but they are clearly ahead of us in understanding the value added of the environment or the media, so that's kind of where we are right now on that.

I'd like to say we were further down the road, but I will tell you that discovery and action in this environment is following Moore's Law more than it's following PBPS. People are actually moving in this area. We're getting value and capability much faster than we get it on the hardware side of the equation, so to speak. I'm very encouraged by what's going on in that environment. But the scale for the most part right now that we're trying to deal with, and the services are just trying to wrestle the value equation out on this.

Question: We've been working [inaudible]. [Inaudible]. [Inaudible] bureaucratic [inaudible].

General Cartwright: Oh, say it's not true. [Laughter].

Question: [Inaudible] situations where a lot of people [inaudible] that [inaudible] warfighters [inaudible] that literally [inaudible]. [Inaudible]. Can you talk about how [inaudible]?

General Cartwright: Sure. All of us would like to paint bureaucracy as evil. It can't respond as quick as we want to make profit or deliver or make decisions, have kinetic output, whatever. It depends on where you sit as to what you really want, but the bureaucracy is always the adversary for the most part until it saves your butt some day and then okay, I can tolerate this.

But the reality here is that there is a discipline that has to be applied to wants, needs, output, resource, et cetera. One of my favorite quips is that, and I'm extremely guilty in that I was in the requirements side if the equation, led the requirements side of the equation, might have had something to do with JSIDS, but when you look at that activity, if there's want absent resource, that's not vision, that's hallucination. You've got to have both. You've got to match it up. You've got to judge it by the goal that you're going to as to how close you're getting to that goal and whether it contributes or not.

For the requirement side of the equation, most things have been defined. A lot of times people are trying to get you to go through bureaucratic things for no good reason. If you go in and work the inside of the system, you [inaudible] this requirement, I can work under this requirement, and I'm okay. And in the area that you're talking about, that's exactly what we've done. We've gone in and said wait a minute, there's a requirement for precision here. I may be dealing closer to the objective than I am to the threshold, but that's all right, I'm still in the boundary space. I can use that. And I can move through this.

There are ways to move through the system. I'm not in favor of just throwing the system out because there is value in it. We've got to have discipline. There's got to be discipline enough to allow competition. So I think it can be worked.

My biggest complaint about the system, the bureaucracy, is that it is telling the user what they need. That business practice went out years ago but it's still alive and well in PBPS and somehow we're got to fix that.

Back to your issue of precision, and it goes back to my discussion about coms. Precision, navigation, timing cannot come from must one source. The kind of work that you're doing and the kind of work that others are doing to allow us to get that absent the availability of GPS and to move it around in ways, different antennas, different wave forms, et cetera. Hugely important. Again, I'm a Marine. I've got to have three ways of coming at any problem. Anything less than that and I'm not going to buy the new toy. I'm going to buy a third way of doing the same thing. I just have to. On the battlefield it's what's going to save your life. So I really do applaud the work that you guys are doing because it's really shown great value.

I'm trying not to speak badly of OSD and the Joint Staff, as you can tell.

Question: What do you see the role of STRATCOM, your vision, as the Air Force moves forward and tries to set up or is setting up a Cyber Command?

General Cartwright: I look at each of their services and their movement toward some articulation of a cyber capability for each of the services. The flash of my favorite light blue service declaring that they're going to move in that direction is good and welcome and needed to occur. It's the fourth service to go there. The Navy has moved in that direction aggressively several years ago and is well down the path. Navy and Army are way out ahead of the other two services right now. That doesn't mean that catch-up is anything like it used to be that can be very quickly done. But what they've got to come to grips with is kind of back to the other question which is at the expense of what? What am I willing to give up to enter into the advantages that this new medium might offer me? Am I willing to give up a squadron of F-16s and take all the people and move them over to cyber? How much is that worth? Is it really worth as much as that squadron of F-16s? That's really what they've got to come to grips with, absent the flash and this is what we can do and the imperative statements that go with it. The worry about intrusion and things like that. Absent that, at the end of the day what resource are you willing to give up in a zero sum game in order to go do this? And does it show sufficient value to compel you to do either more or less of what you need to do? That's the question for the services.

They'd all like to see somebody else jump in with a lot more early to see if it works and then come in behind, but they know that if they're going to get the competitive advantage they've got to stick their nose out there in front, so it's a huge conundrum for them. The good news is, all of the services are now convinced that this medium has some promise. What they're trying to decide is how much and then how heavily do they want to invest.

That's the simple side of it. You know it's more complex than the cultures and everything else in that, but that's really at the end of the day where the Service Chief and Service Secretary's head has to be, is how much of this zero sum game am I going to take out of something else and put into this area?

Question: I wondered if you could spend a minute talking about the movement of the kill chain to the kill web and [inaudible] validation for these operational tests and evaluation [inaudible].

General Cartwright: The flexibility that you get moving from a kill chain to a kill web is hard to argue with. That's probably the biggest draw feature to the activity is that you can do mixing and matching in ways that you never thought you could do in the past, so you can afford both ad hoc coming together of distributed forces, whether they be US or allies and US which is really what you're trying to advantage, in ways that you never thought possible and give them capability you never thought they could have as small entities. That's the draw.

The question here is how do you validate the whole end to end activity and how many of these do you want to pay for, and how many of them do you want to have what I would call time to market on. In other words, I've got enough information that if I found value there in a month, in a week, in a year, whatever is appropriate, I could connect all those dots. How many of those do you want to have as kind of hedges against uncertainty? Then how do you turn that over to the testers and say here's what I want to be able to demonstrate. Today in missile defense we do it through the engagement sequence group so we'll do the connection of a ground-based interceptor to a Cobra [inaudible] terrestrial radar, to an Aegis [inaudible] platform and we'll connect all those dots, do an actual firing, and demonstrate that the quality of service is sufficient to make a probability of success of X. Eventually these webs get so intricate that you start to worry about either stack up of tolerances or the ability to actually create ad hoc relationships that were empowered in ways you never even thought of when you put the network together. We haven't gotten to either one of those yet, but that's something that's easily within the next year that we're going to start to bump up against.

The same goes in ISR, by the way. ISR uses this and we're trying to get this into the lexicon and let them start to understand how you can do this because this is your classic high demand/low density activity, so relationships are critical. What we really are trying to get to in the ISR side of the equation is kind of an in the kill web, an awareness of who you are and where you are, an identity. If I know the identity of the platforms, then if I, let's just take a geo-location scenario. If I've got an overhead asset that can sense a signal and it knows where it is, and because it knows where it is and everything else knows where it is, so to speak, then if there's an airplane and a ground unit in that footprint, they know they're in a footprint, they can create an automatic relationship and they can geo-locate a signal very quickly. That's where your recon flight could be, that's where your special forces would love to be, that's the kind of relationship that starts to get this web, but those are ad hoc relationships that are built on demand based on an identity and a knowledge of where you are and what you can do. We're nowhere near that sophisticated yet, but that's where we're trying to push.

So we're running the ACTDs, et cetera to start to demonstrate that, but then what are the critical testing parameters that you go after so you don't have to do what we do today with airplanes which is every time I change the OFP I've got to go out and fire every single weapon. It just drives you crazy. It adds at lest a half a billion dollars for almost every OFP upgrade.

So I want to change two lines of code, and of course my, you shouldn't start me on this, my OFPs, you pick them. You change a line of code in the thing, it costs you an arm and a leg, takes three to five years to do and a huge test budget to boot when shoot, when I sit down at my computer and plug in a printer it knows it's a printer, goes and finds the driver. The architecture was built the right way. Why can't we start changing the architecture in our military system. Then we're like, you pick the company.

That's what's out there in front of us. That's the big challenge on the kill web side of the equation.

Thank you very much.


Find this speech at:
http://www.stratcom.mil/s6