DCI Speech 05/22/96

ARPATech '96 18th S&T Symposium, Atlanta, Georgia
"The Future of the National Reconnaissance Program"
John Deutch, Director of Central Intelligence

I am most honored to have been asked to address the 18th annual ARPA Science and Technology Symposium. Like almost every technical person in this country, I have some familiarity and considerable respect for the contribution that ARPA has made to advancing technology in the United States and thereby profoundly strengthening our national security and our economic strength.

Moreover ARPA has developed a unique way of managing the innovation process, one that is widely admired but rarely duplicated by other government agencies, by industry or by other countries. It is a unique approach that places great reliance on individual program managers and their ability to smell out promising new technological ideas and work with industry and military users to realize innovation.

This has been a formula for success for many decades. It has brought advances in a wide range of fields: computers, robotics, materials science, electronics -- a roster almost identical with those areas where the US has its comparative technology advantage in both commerce and defense relative to other countries.

And what is so notable about this success is that ARPA has been able to adapt to changing circumstances and to innovate. ARPA has had to be willing and able to drop old technologies and pursue new technical opportunities to respond to new national security needs. It is a remarkable achievement that should make everyone who is or has been associated with this organization very proud.

I am confident that ARPA's success will continue in the future. It is my strong impression that ARPA has its technical eye right on the ball and that under my friend Larry Lynn's leadership, it will continue to make important contributions.


I want to spend a few moments discussing with you the challenges facing another great US technical organization -- the National Reconnaissance Office. The NRO has had a very distinguished, but very different history from ARPA. The NRO has made tremendously significant contributions in an area that has been of great importance to this country -- developing satellite systems to collect imagery and signals intelligence from space for both national and military users.

The mission of the NRO was important during the Cold War; the mission remains important today. Policymakers demand global technical intelligence to understand a tremendously wide range of subjects. Let me mention a few examples: ethnic cleansing in Bosnia, the extent of coca cultivation in South America, compliance with arms control treaties, and intelligence and warning of military threats around the world, for example, the recent Taiwan Straits Crisis.

The Gulf War made clear the tremendous potential of technical intelligence in support of military operations. Many knowledgeable observers believe, and I concur, that the intelligence provided to our joint military commanders to support accurate delivery of precision munitions will form a foundation for our future military superiority. This technical intelligence will be provided largely, but not exclusively, by satellite systems developed and acquired by the NRO.

The question is how do we organize the Intelligence Community to meet this future challenge. "How can we best use the traditional strength of the NRO system"? "Can the NRO adapt to the changed circumstances of the post Cold War world"? Or perhaps even more precisely, "Will the institutions responsible for governing the NRO -- the Department of Defense, the Intelligence Community, the Congress -- encourage or stifle the adaptation process"?

These are important questions, because technical intelligence is vital to our future strength both for policy and military matters. If the NRO organization as we know it, is going to prove to be incapable of efficiently meeting future intelligence needs, we are going to have devise new mechanisms for doing so.

In order to examine these questions we must take a look at the origin of the NRO and how it has traditionally carried out its mission and compare that experience with today's circumstances.


The NRO was established in great secrecy in the early sixties for the purpose of building intelligence satellites. The earliest work on intelligence satellites, the famous CORONA program, began at the CIA at the urging of technical experts, notably Din Land of Polaroid and Jim Killian of MIT, who recognized the technical possibility, at that time largely unimaginable, of taking pictures from space.

This origin led to several still essential elements of the NRO. First, it is an organization that is technology driven to provide system solutions to satellite intelligence possibilities. From its beginnings, the NRO program focused on getting the highest performance from available technology for users of its products; often cost was a secondary consideration. The assumption was that because satellite intelligence products were so unique in providing critical information that the military and national users would pay whatever the cost.

Second, and appropriately for the time, the NRO focused on collection technology; less attention was given to exploitation and distribution of product or to analytic tools necessary to fuse the different types of technical intelligence.

Third, the "technology push" perspective, combined with the security constraints of the Cold War meant that the NRO did not develop mechanisms to bring all of the users, principally the military users, into a decision process that would balance needs, technology opportunities, and costs.

Fourth, the importance of the mission combined with the highly secret nature of the NRO led to a unique procurement process, largely free of the usual cumbersome government acquisition oversight process. Over the past three decades, the NRO has demonstrated the advantage of this flexible acquisition process in getting the job done. But, we must also assure that there is adequate financial oversight to avoid situations as we have recently seen, in the public discussion of unreasonably large NRO forward funding balances.

Finally, in part because of its origins and also because our satellites serve both national and military users, the NRO is a hybrid organization. It reports to the Secretary of Defense, but it receives much of its program guidance and collection priorities from the Director of Central Intelligence. The staff of the NRO are employees of the CIA, the Air Force, Navy, National Security Agency and other DOD elements.

Over the past three decades the NRO has been an enormously successful technical organization. No one can deny that it has produced satellite systems unequaled by any other nation. Intelligence provided across peacetime and wartime has been crucial to our national leaders. It is a success story that rivals ARPA's. The individuals who have served in the NRO should be proud of their accomplishments, and all Americans should appreciate the contribution this great organization has made.


However, both the end of the Cold War and the pace of recent technological change present the NRO with very different challenges. I am concerned that both the NRO organization and those who are responsible for oversight of the NRO - - in Congress, the Department of Defense, and the Intelligence Community -- understand the changed nature of these challenges and that they are prepared to give the NRO the support required for continued success.

The two key changes that have occurred concern changing technology and the growing importance of resource constraints.

With respect to technology, technical collection is no longer constraining; many attractive technology options have been largely mastered; the problem has become to select those to exploit.

The information revolution has shifted the focus from collection to the process of efficiently processing, exploiting, and communicating data collected from space, airborne or other platforms to end users of intelligence. Of course, over the years the NRO has recognized the growing importance of processing and exploitation. But, it is only recently that the centrality of information processing has been appreciated and the significance for NRO operations acknowledged.

For example, in the area of imagery, the possibility now exists of managing the data flowing from the photons that hit the focal plane on satellite (or other platform) sensors to produce near real time imagery products, whether maps or pictures, tailored to user needs. We no longer need separate imagery exploitation and mapping activities, but rather we must develop a single process to manage a complex, geolocational information system. As I mentioned, this transition is not discretionary; we must do it if we are to enable the advantages provided by a flexible geospatial information data base, on-line with our military users.

The implications for the NRO are significant. First, information demand pull must replace satellite collection technology pull. Users must have the dominant voice in determining what collection is needed.

Second, satellite architecture must be seen as embedded in a larger information system. This system can be configured seemlessly to meet user needs including exploitation and communication. We have taken certain steps to facilitate this. Bill Perry and I have formed the Joint Space Management Board, jointly chaired by the Under Secretary of Defense for Acquisition and Technology and the Deputy Director of Central Intelligence, to assure a single corporate approach to both military and intelligence space matters, including, for example, space launch needs, ground architecture, and the interaction between space communications and collection.

We have also jointly recommended to Congress the formation of a National Imagery and Mapping Agency (NIMA) to assure that a single entity is responsible both for exploitation and distribution of data collected from both space and airborne imagery platforms. NIMA should bear the same relationship to the NRO in imagery that NSA does in SIGINT; NIMA and NSA are principal NRO product customers responsible for data exploitation.

There are other implications of this change in technological focus. Let me mention a couple of examples. In the post Cold War world, we can expect that US forces will be deployed for varied purposes in many parts of the world, just as they are today deployed in Bosnia. The rule, rather than the exception, will be that they will be deployed with coalition partners, notably NATO members. For maximum effectiveness, we will need inter-operability with NATO and other coalition partners.

Therefore, we must anticipate sharing more intelligence information with our allies. This, in turn, leads to greater interest in cooperative development, acquisition, and operation of C3I systems, including satellite based intelligence within NATO. It is both in our strategic interest of strengthening the alliance and in US industry's interest to move in this direction.

From both the point of view of alliance unity and economy it will be a great misfortune if NATO decides to proceed toward independent satellite reconnaissance and military communications capabilities. I believe that Europe will not be willing or able to devote the resources necessary to accumulate the kind of capability that we have. We therefore should be willing to move to greater sharing of technology and production to provide NATO the incentive to join us in cooperative trans-Atlantic partnerships. Such cooperation would be a major departure from past practice of the NRO and over the long term will significantly strengthen our common industrial base.

The second implication I want to mention is that technology advances now allow commercial suppliers, rather than the traditional NRO custom hardware and software contractor, to provide a greater fraction of intelligence and reconnaissance needs.

At one level, this means that the NRO, like all other parts of the defense establishment, should seek to rely more on commercial products and services rather than government unique items, because, especially in the areas of information technology and electronics, it is cheaper. But, at a second and potentially more important level, we should be prepared to consider relying on the commercial sector to provide directly some of our information needs, whether maps, pictures, or communications service.

I have mentioned the important technology changes to which the NRO must adapt. It is equally important that the NRO adapt to the changed circumstances affecting resource availability.

The time has long past when costs could be considered secondary in planning and executing the NRO program. The Cold War is over and understandably resources will be tighter for both defense and intelligence. Today, we have an abundance of exciting and important technical collection opportunities and the problem is to make choices among them.

The choice of which competing systems to pursue depends upon comparing the performance, technological risk, and cost of various satellite system alternatives. The critical user is the Department of Defense because technical intelligence is so important to support military operations. When we buy a new satellite system or acquire a enhanced capability it usually is to meet a predominant military need and it usually is paid for largely from the Defense budget. This effectively trades-off additional military platforms for more intelligence. Therefore the key judgment in the decision is whether the military capability to perform a mission is better achieved by buying additional intelligence or more platforms. Accordingly, the needs process must involve in a central way the ultimate user of intelligence and not just intelligence specialists or intelligence system developers, e.g. the NRO.

Defense is not the sole user of satellite capabilities. The Department of State, the National Security Council, the CIA (to name just a few) are also important customers for the products provided by the NRO. Thus, while military needs drive NRO systems, the particular needs of other users must be accommodated in establishing the architecture.

We have in place a good system for making these judgments. For the past four years the Deputy Secretary of Defense and the Director of Central Intelligence have been co-chairing a board composed of the leadership from the DOD and the Intelligence Community to consider the choices.

We need to improve continually the analytic basis on which decisions are made. The excellent NRO technical analysis that assesses performance measures such as area, quality, revisit time, and assured access, are not sufficient. The analysis must reach to an assessment of the contribution of intelligence systems to military outcomes in scenarios that are judged to be consequential. Recent sensor to shooter analyses are examples of this approach.

I trust this discussion adequately describes the tremendously changed world that the NRO faces. Let me summarize, perhaps overly starkly, the tremendous change the NRO faces from the past to the present:

Technology pushUser demand pull
Satellite collectionInformation distribution
Collection constrainedExploitation central
Cost secondaryResource choice primary

Clearly, the NRO is being asked to adapt to a lot of change. How are they doing at meeting these challenges?


I believe that the substance of the NRO program is responsive to the changes discussed above. The basic thrust of the intelligence satellite architecture, initially sketched by an outside panel chaired by Jim Woolsey in 1992, before he became DCI, pointed the way to a capability that represents the first step toward giving the warfighter dominant battlefield awareness and providing timely intelligence to national level policymakers. The program is affordable and the recent changes introduced by the Deputy Secretary of Defense and I assure that there will be strong financial management.

I cannot discuss the details of the satellite architecture which is being pursued. Let me simply characterize it as a technical step forward that increases needed collection capability and that it responds to prioritized user needs. The program includes attention to ground architecture for product distribution. The architecture includes significant change from existing satellite systems and therefore demands scrupulous attention to implementation of the program.

There are also many visions of what the next step should be for our satellite architecture. Its often tempting to move to the next generation before the present generation has been successfully implemented and reached the end of useful life. The danger is that we take on too much technical risk and that we embark on a program that exceeds likely available resources.

I am concerned that those outside the Intelligence and Defense Communities do not appreciate the importance of continuing the current program before moving to the next generation of technology. There needs to be a stronger appreciation of the time it takes to execute a program and the importance of sustained technical attention to program success. Moreover, effective management requires a predictable stream of resources. It simply is not possible to have an effective program if major program changes are introduced in every budget cycle.

Let me conclude by answering the three questions that I posed earlier: First, we can use the traditional strengths of the National Reconnaissance Office and its dedicated workforce; it would be foolish and wasteful not to make use of this organization's strength. We need to continue the technical advance and leverage the commercial marketplace in the areas of sensors, communications and large scale data handling.

Second, the NRO must adapt to change by moving toward an integrated imagery system that places greater emphasis on exploitation and distribution of product to military and national users. As the focus moves to data exploitation it will radically change the way the NRO and other intelligence agencies do their business. NIMA is one bold step in this direction.

Third, the overseers of the National Reconnaissance program -- the Department of Defense, the Intelligence Community, and Congress -- must recognize that these important technical changes will only come gradually and only according to a stable and predictable multiyear program.