Mind the Gap

CHAPTER FOUR

Creating a Transatlantic System of Systems
The Third Tier

Closing the gap will require two types of efforts to make forces compatible. European forces, as the last chapter argues, must outfit and reform themselves to be more like U.S. forces; by doing so, both we and they can undertake RMA missions in complementary fashion. As this chapter argues, U.S. and European forces must also be linked together, so that both sides can share a common knowledge of the battlespace.

Working Together

The need for compatible forces places most of the work on the Europeans. Most of the work needed for information systems compatibility, though, lies on the western side of the Atlantic. The United States is well ahead of Europe in integrating its various sensors, networks, databases, and weapons into what retired admiral William Owens has called a "System of Systems." U.S. decisions made in the process of this integration will color similar decisions made on behalf of NATO as a whole. If the United States constructs its System of Systems for itself alone, rather than for the alliance, the prospect for a NATO projection capability will be bleak, even if the Europeans build more suitable forces.

In other words the United States owes the alliance the obligation to build a plug-compatible System of Systems; Europe, in turn, has the less onerous task of buying the plugs and making intelligent use of the data it receives as a result.

Building on the last chapter's force requirements for a transatlantic RMA, this chapter discusses the following:

Creating the System of Systems and the "Grid"

It is important to understand what the most serious discrepancy is between European and American military technology. The United States, as much as any country, and more than most, has acted on the premise that the way to improve weapons systems has been to inject successively higher doses of technology into them. Today's tank is more capable than yesterday's: it is faster, tougher, shoots more projectiles faster and more accurately, has more sensors, and is generally more intelligent. So, by and large, are today's ship, today's airplane, and so on. Tomorrow's will be more capable yet. But we are used to constant progress and, indeed, expect it. Progress, per se, is not revolutionary. The United States, by constantly buying hardware, has technologically better equipment than do the Europeans, on average. This was true in the Cold War; it remains true today. But this difference did not keep NATO from functioning well those many years. Hardware is not the gap that most needs closing if NATO military effectiveness is to be regained for the new era.

What is revolutionary is not the caliber of equipment, but the fact that information technology in general, and data networking in particular, is giving rise to an American military that is much more than the sum of its parts. The conjunction of sensors, networks, databases, and weapons--a System of Systems--has the potential for radical improvements in conventional military operations.

Sensor fusion may serve to show how networking can help. Battlespaces are likely to grow more transparent to the United States, not just because sensors are better, but because each individual sensor is part of a network of space-based assets, aircraft, unmanned aerial vehicles, naval radars, ground sensors, and human observers. Sensors using multiple technologies (e.g., electro-optical, infrared, millimeter-wave, radar reflection, acoustic) can collectively illuminate the battlespace much better than each might individually. Sensors that can distinguish location accurately but that are unable to identify objects are not fully useful; nor are those that can identify objects but that have little ability to locate them. By uniting these sensor functions, RMA forces can understand the battlefield. If the movement of a target can be given in real time to weapons that guide themselves by reference to a moving dot on a map, the combination offers a highly lethal capability against any visible target.

Good illumination by itself cannot foster the high-tempo RMA operations described in the previous chapter. Information is decisive only when it can be analyzed rapidly, converted into plans of action, deconflicted with all other ongoing actions, resourced properly (e.g., logistics), and effectively communicated to operational units. Widespread networking fosters faster and often better command and control. It permits operational planning to be conducted more effectively and the results distributed farther in less time. It permits greater synchronization of forces in action, and improved coordination in the face of fleeting opportunity or unexpected danger. Forces out of sight from each other can maneuver more efficiently. Augmenting voice communications with video, imagery, and databases, notably maps, fosters an integration of effort that hitherto required face-to-face contact.

Enhanced speed counts. Take the challenges of hunting for SCUD launchers, picking out a missile-bearing speedboat in a crowded harbor, or conducting urban warfare. All three depend on the ability to sense a rapidly emerging threat, determine its characteristics, and respond effectively--within minutes. To the extent that the WMD threat requires friendly forces to be physically dispersed, coherence is reduced. Networking can integrate them and overcome the effects of dispersion.

A System of Systems also helps with ancillary functions: keeping track of logistics, coordinating repair cycles, conducting what-if exercises, modeling the battlespace and simulating alternative operational concepts on it, distributing training materials, and gathering the operational experience of everyone (e.g., after-action reports) to build a knowledge base for all.

Today, networking is an advantage. Tomorrow, it may be a necessity. Most of the basic information technologies behind the RMA arise in the commercial world. They may therefore be purchased by anyone with the means to pay--and prices are declining. Even a Third World military can be greatly improved by acquiring and integrating laptop computers, cellular telephones, digital videocameras, wireless microphones, pocket radars and laser rangefinders, GPS receivers, turnkey air traffic control systems, and services from space. Those who can absorb such technologies could see us better. The RMA can make the battlespace transparent for everyone, and thus untenable for the bulky platforms with which 20th century militaries still fight. JSTARS and AWACS, for instance, are aircraft without peer as sensors; but neither is particularly stealthy, and, to function, both must emit copious quantities of energy, making them quite visible. Survival favors having many small things over a few large ones, but making the many work as well as the few requires they be networked.

The United States is networking its forces and sensors today. Should it one day need to disaggregate C4ISR platforms into networks of small and inexpensive sensors, and its strike forces into networks of small elite units, this transition, though difficult, will be possible. If Europeans have not caught up, they may not make the transition at all. U.S. and European forces will be distinguished not simply by efficiency, but by the very ability to function in such an environment.

Created by the conjunction of sensors, networks, databases, and weapons lies what may be called the Grid: the virtual information system that ties these systems together, and makes them collectively accessible to operators. The Grid would be the means by which devices and databases could be accessed, and upon which services and applications reside. As users of the Internet and the World Wide Web intuitively understand, virtual entities are very real.

Being "on the Grid" continuously and effortlessly should become second nature for U.S. forces of tomorrow. By so doing, warfighters share a real-time map of the battlefield, annotate this map for others, find out where parts are in their repair cycle, participate in a simulation or exercise, assess the state of the network (and perhaps defend it from attack), diagnose remote equipment, and call for fire support from remote weapons.

The distinction between the Grid and the System of Systems is of great importance for the prospects of the NATO RMA. For Europeans to create their own system of systems of comparable capability, they would need to invest in component systems--platforms, weapons, and sensors--to the same degree and of the same kind as the United States does. For the Europeans, however, to join the System of Systems, and therefore be on the Grid, would be easier, requiring only acquisition of compatible hardware and software. Indeed, exploiting the RMA as an alliance makes being on the Grid essential. Otherwise, in military operations in which U.S. forces are on the Grid, and European forces are less (or not at) all connected, the latter's contribution may be not only wasted, but counterproductive--even if they have RMA-type forces, doctrine, and training. In other words, investments by the allies in better forces would be in vain, from an alliance standpoint, if U.S. forces are on the Grid and allied forces are not.

The consequences of disjunction can be illustrated by a hypothetical vignette. Assume U.S. forces, using the Grid (with its sensor fusion, automatic target recognition, and template matching capabilities), have spotted and confirmed enemy armored vehicles moving out from cover. Such vehicles may not linger exposed forever, but a few minutes should be long enough to strike them with ground-based standoff weapons. On their own, U.S. forces could launch such strikes, but European forces are working the area. Therein lies a dilemma. U.S. forces are uncertain where Europeans are working (e.g., because they do not report their locations automatically). Indeed, enemy forces may have moved precisely because European actions had flushed them out. Launching a standoff strike without checking may frustrate an ally's plans (which, having been made up on the spot, were not reliably and rapidly linked to U.S. operational plans). U.S. and European forces may not even be fielding compatible IFF (identify-friend-or-foe) systems. U.S. forces may want to send the location, bearing, and signatures of the adversary vehicles to the Europeans so that they can target enemy armor precisely. But poor network interfaces, limited bandwidth on the receiving end, and translation difficulties between U.S. and allied systems would cause delays. Allied forces might also want to drill down into the surface data to understand the rationale and thus validity of this intelligence, but there is no way to pose a query from one national intelligence system to another. Prudence dictates that U.S. forces hold off striking what they can easily see, and opportunities are continuously lost.15

Those on the Grid have a better opportunity to see fleeting opportunities in near real-time, pass critical information throughout the ranks, get instant battle damage assessment on strikes and other operations, and feed new conditions into ongoing plans. This information will enable the U.S. military to take the initiative, conduct operations, assess their effects, and generate new options faster than the other side can react. The Grid also gives U.S. forces the ability to act at places of their own choosing. With a growing ability to project sensors across the breadth and depth of the battlefield--from space, unmanned aerial vehicles, or scattered on the ground--coupled with weapons that can exploit this information, the U.S. military is able to strike deep into hostile territory.

In sum, if U.S. and allied forces are to fight together effectively, a NATO Grid, and nothing less, is imperative. Even if the United States alone contributes certain assets--space surveillance, long-range UAVs, direct broadcast satellite connections--to the Grid, they would be operated on behalf of all. Put another way, if the Europeans are to adopt the U.S. emphasis on rapid operations and deep strike, they need to have access to information and networking capabilities that enable such operations to take place.

Current Issues

Unfortunately, getting the allies on the Grid will not be as easy as, say, giving them an Internet account or the address of a Web site. Under military conditions, establishing networks is not trivial, addressing security concerns is vexing, and common command-and-control arrangements and practices to support RMA operations must be worked out. Moreover, many differences exist between the receipt of data and the generation of operationally useful knowledge. Because the U.S. Grid is "under construction," the salience of these issues is not yet apparent. Yet, their emergence can be perceived and their importance extrapolated by looking at a few current issues: (1) Bosnian operations which suggest problems in forging compatible C4ISR links between U.S. and European forces. (2) Theater missile defense (TMD), a potential show-stopper in the Persian Gulf, and, soon, the Mediterranean. (3) The conduct of U.S. experiments which may lead to policy decisions that will shape the Grid for years to come.

NATO in Bosnia. In late 1995, IFOR (Implementation Force), a multinational coalition under NATO command, deployed several divisions to Bosnia to implement the Dayton peace accords. For the purposes of ground command and control, Bosnia was divided into three sectors: one for the United States, one for Britain, and one for France. Each, in turn, commanded forces from many nations, but mainly NATO allies and "partners" (e.g., Ukraine and Russia).

Bosnian operations have been generally successful, and a C4ISR infrastructure was able to support them adequately, but not without a good deal of jury-rigging to link incompatible systems.16 Establishing interoperable communications required NATO to build a system from scratch out of heteroge-neous elements in a region they had not planned for, and in terrain that complicated line-of-sight communications. Nevertheless, C4ISR was established despite the shortage of key connectors, shortfalls in satellite capacity, the remoteness of supply centers, wide variations in component systems, and constant entreaties to field untested concepts.

The result was a patchwork. Numerous overlapping communications systems were needed because voice, video-teleconferencing, data, and commercial Internet traffic ran over different systems. These overlapping systems also were needed because of differences in the security classification of material (NATO had a great shortage of certified communications security [COMSEC] equipment) and because high-bandwidth satellite systems do not necessarily mix well with low-bandwidth tactical systems. Early in the existence of IFOR, upwards of 20 percent of all voice calls did not go through. Even later on, users had to carry complex dialing plans on 3x5 cards. Things would have been worse had NATO not started working six months earlier on specific communications interface guidelines for Bosnia. The standard NATO interface for analog voice networks is slow, inefficient, and lacks some functions. But it exists, which cannot be said for digital telephone networks, much less data networks, which are far more crucial to RMA operations than to peacekeeping. NATO, at least, had settled on a common software suite, Microsoft Office (divergent software plagued the 1993 Somalia operation).

The severe intelligence mismatch that might have been expected between U.S. high-technology systems and the less-capable systems in European hands was less evident in Bosnia. The former were developed for conventional high-intensity war and had to be heavily adapted to fit operating conditions that put a premium on hand-crafted databases to record day-to-day activities. Signals intelligence and overhead surveillance from UAVs helped build the overall picture,17 but the foundation for understanding Bosnia was human intelligence. Here, the British experience in dealing with Northern Ireland proved useful in prescribing collection methods for Bosnia. Indeed, Europeans often grasped the nuances of the Balkans more readily than did the Americans.

A roughly common IFOR intelligence picture came into focus even though contributing nations differed sharply in their practice. Americans were generous in sharing what they knew in order to build common situational awareness, while the British employed strict need-to-know criteria. One NATO command-and-control system (CRONOS) was brought into theater, and another

(LOCE) was used to disseminate intelligence; but there was no electronic interface to equivalent British systems or U.S. national and strategic and tactical data networks (hand-carried information was the substitute). The NATO above-secret message network, TARE, had a node in IFOR headquarters, but was extended no further into Bosnia.

The common air picture, which necessarily included commercial aviation, is where integration made the greatest strides in Bosnia. NATO had a common air operations center to administer the reconnaissance assets of the 16th Air Force, whose vice commander, in turn, ran the center. Even so, barriers prevented the common air picture from being fed into the common operation picture (COP) hosted on the U.S. Global Command and Control System.

The ability of U.S., British, and French forces to work together in Bosnia, under the IFOR structure, but using NATO and national assets, suggests that a NATO Grid is feasible. Success in Bosnia would have supplied better proof, however, if the situation had been more stressful. The jerry-built system never was tested in battle conditions, and it might not have withstood the stress from bursts of communications loads if combat had erupted. Peace forces faced fewer urgent response requirements than if they had been fighting high-intensity engagements at blistering speeds. Bosnia was neatly divided into three segments, and there were few requirements for combined military operations--none where smooth interoperations might decide between victory and defeat. The air picture was relatively benign, i.e., lacking air or missile attacks on NATO forces. Conversely, the security problems in Bosnia that arose from having to share intelligence with less-than-firm allies (e.g., Russians, Moroccans) may not necessarily be a feature of high-intensity combat.

U.S. planners speak of tomorrow's Grid as being able to illuminate a battlefield the size of Ohio, pick out and track every military target of relevance in real-time, and give all units and commanders whatever information they need. No such U.S. system exists today; certainly not in Bosnia. According to doctrine, NATO commanders in peacetime must rely largely on member states for their needs. In Bosnia, this was good enough; for combat, against a determined and sufficiently equipped foe, perhaps armed with WMD, it is clearly not.

Theater Missile Defense (TMD). The United States is spending several billion dollars a year devising protection against hostile missiles. Because the Europeans spend much less, this is an area where differences in viewpoints widen the transatlantic gap. In time, given the increased salience of the WMD threat from rogue states, one would expect European and American views to converge.

Even if TMD investment were comparable, defending forces and populations against theater missiles soon may require the United States and European nations to meld at least part of their real-time intelligence systems soon.

A viable TMD system that covers the European allies must be part of an integrated NATO TMD system, which, in turn, would have to be on the NATO Grid.

The need for coordination comes from the complexities of missile defense. The current U.S. concept envisions four tiers: (1) direct attacks on missile sites, (2) airborne lasers to knock out missiles as they enter the stratosphere, (3) upper tier missiles for long-range engagement, and (4) lower tier missiles for close engagements. If the first three worked, a U.S. TMD system could defend allies over a wide area. But the first three do not suffice today--despite operations against SCUD sites, none was validated as destroyed by air attacks in the Gulf War; the airborne laser is untested; and the Army's system (THAAD) has failed field tests. Thus, short-range interceptors are still needed, and U.S. short-range systems would not be likely to protect European forces (much less European cities). The United States will share with Europeans its anti-missile technology but it is hard to see the United States dedicating part of its own inventory to defend Europeans forces or populations when Europeans could easily acquire their own assets.

So, Europeans must buy their own short-range protection. But they must have information if such systems are to work well. Europeans insist that the United States supply them early-warning data on missile launches. Because this data is so sensitive, progress in sharing has been slow. The new U.S. space-based infrared (SBIR) constellation will generate even more information, data that can help with missile interception. But the challenge does not stop there. Lower-tier defenses alone may not work well, and they should be considered but one layer of a four-layer defense. Engagement systems, in turn, are but one part of a complex antimissile architecture that includes early-warning sensors, peripheral surveillance systems, threat processors, and civil defense (including NBC detectors). At the very least, complex command-and-control and hand-off arrangements will have to be devised. In this case, a NATO Grid is indispensable for combined TMD operations, which, in turn, will be indispensable for alliance military operations, e.g., against WMD-armed rogues.

Experiments. Experimentation with advanced C4ISR systems and RMA doctrine has become an important step in reshaping U.S. forces. Again, the role of the Grid looms large. The Navy is networking its ships through a Cooperative Engagement Capability (CEC). The Army is digitizing itself through its Force XXI program. The Air Force is beginning to develop powerful methodologies for global operations planning. These experiments offer some broad hints as to whether and how the United States will draw allies into the new operational concepts debate.

The Navy's CEC network enables ships in a battle group to combine their radar data to better defend against cruise and ballistic missiles. Cooperation permits ships, acting together, to determine the track of incoming missiles more accurately than any could have done by working alone; it also lets any ship, regardless of its own radar picture, engage incoming warheads. The CEC was initially considered such advanced technology that it could not be shared with allies. But this meant that the contribution of European ships to the defense of any combined naval task force would have been zero if the network were employed only by the United States. Recently, the British Navy was brought into the net, with others presumably to follow.

The Army's Force XXI program aims to digitize ground forces by providing every vehicle with a node on a tactical internet. Thus linked, each vehicle could broadcast its precise location, share a common electronic map of the battlefield, and distribute mission orders and related data. Although the U.S. Army intends to digitize all its divisions no later than 2010, it has yet to begin serious interoperability testing with the U.S. Marines or the Air Force--much less with allies. The Germans, who are investing $10 million a year in the process, are active observers and, to some extent, participants. British forces are following events closely. The French and Americans have been looking at several experiments to determine how their various weapons platforms can transfer data.

Work is underway to develop a set of commonly accepted standards, protocols, and message formats that allies might use to coordinate actions. By and large, however, NATO allies are willing to let the United States take the lead, learn from its successes and failures, and, if worthwhile, invest later. But this strategy runs the risk that, once developed, the details of digitization will be so deeply embedded in the U.S. military that they will be difficult to change to accommodate Europeans. For the United States then to say the Europeans must comply (or stay home) will be hardly ideal from the perspective of combined operations.

These experiments suggest that a difficult trade-off awaits Grid designers. Information systems exist to inform operations, and their design reflects tacit and implicit assumptions about how warfighters do their job. Both the assumptions and their effect on systems design become obvious only when warfighters with a different set of assumptions begin to use such systems. Even if systems design is consciously associated with a particular set of practices, a placeholder can be left to accommodate other, equally valid allied practices. But the later the Europeans are brought into the process, the more deeply etched the details will be and thus the harder to change. And, as anyone working on the Year 2000 problem can attest, details matter.

Principles for a NATO Grid

A NATO Grid would be to NATO RMA forces what the U.S. Grid is to U.S. RMA forces: the means by which the battlespace is illuminated, forces networked, command-and-control integrated, and ancillary functions (e.g., logistics, training) made more efficient. A NATO Grid should be viewed as a utility that any alliance member could use if it had plug-compatible equipment such as receivers, workstations, fire-control systems and targeting modules, and, of course, software. Europeans already own digital hardware, and some have the rudiments of their own Grid in place or under construction.

In a NATO Grid, ownership of components--such as sensors, switches, processors, or knowledge-bases--ought to matter less than such features as reliability, performance, accessibility, interoperability, and security. Data from a British UAV electro-optical sensor can be linked through a U.S. network to readings from Dutch microphones so that the bitstreams can be fused with the help of a French-hosted software agent and compared to a German-provided database of marine templates to provide targeting guidance to a topside gun on an Italian frigate. A NATO Grid could include civilian elements (e.g., air-traffic control tracks of European Union countries, NASA-hosted image deconvolution software) and commercial elements (e.g., CNN footage of a harbor area, shipping manifest databases privately maintained for fee). Although coalition partners may believe their own assets first, the Grid can be designed so that technology does not foreclose using assets of others.

The most important issue entailed in building the NATO Grid--especially for the European militaries--is how integrated, thus how interdependent, the Atlantic militaries are willing to be. A NATO Grid constructed by stitching together the U.S. Grid and those of Europeans permits each side a measure of independence. A truly integrated NATO Grid would work far better and be more cost-effective, but it would come with some sacrifice of independence. Since a U.S. Grid is likely to be well underway before a NATO Grid sets sail, the United States is likely to retain great freedom of action. But Europeans have choices to make.

The Germans, for instance, have debated the need to join the French in developing a European surveillance satellite, even though any high-intensity combat is likely to see them operate with backing from U.S. satellites. If satellites become tightly integrated with UAV and ground sensors, and sensors, in turn, are tightly coupled to weapons, then loss of access to U.S. satellites may mean more than simply not getting the pictures they came to expect: other sensor systems and sensor-to-shooter links would be degraded. In practice, European military units will retain whatever sensors, networks, and databases are associated with their current and planned weapons systems. But will those sensors, networks and databases work through the NATO Grid or, instead, will data from an essentially U.S. Grid be but one more source of data that would feed each ally's military information networks? The answer may depend on how fast European nations integrate their own sensors, networks, and databases.

Integration is not the only question facing defense policymakers on both sides of the Atlantic. If European militaries come to rely on a NATO Grid, will they continue to invest in new sensors, networks, and computer programs, or will they increasingly rely on the NATO (mainly U.S.-supplied) capabilities? Conversely, will the United States be willing to provide its allies with local networks and sensors (or sufficient access to global ones) so that the Grid's coverage is as good for allies as it is for itself? If the United States retains complete freedom to operate without European participation, might the Europeans insist on at least some ability to operate without U.S. help, and thus without U.S. consent?

A great deal hinges on how much allies wish to spend for battlespace information that the United States could otherwise provide for little or no cost.18 Europeans may understand that the U.S. RMA may not make them worse off directly, but failure to invest in autonomous capabilities may leave them dependent on the U.S. Grid and thus unable to wage intense warfare on their own even with improved forces. Alternatively, Europeans may figure that they will fight alongside the United States in any big and intense conflict, and that they can afford to conduct peace operations or small-scale expeditions (e.g., into Africa) with pre-RMA forces. However, the more Europe adds to a NATO Grid the greater its effective influence on decisions over its architecture and use.

The reverse also is true. The more U.S. and European strategic and operational perspectives are aligned, the more often U.S. and allied forces will operate together and thus the greater value to the United States of being able to share information with Europeans. It therefore makes sense to design the U.S. or European Grid to maximize cooperation with allies, while also ensuring it can function well without them. Similarly, the more Americans and Europeans trust each other's intentions, the more they can count on the other's forces and the more freely they can share information, satisfied that information will not only be used correctly, but protected adequately.

Mutual confidence affects several issues that affect a NATO Grid:

These are not merely technical issues. They are affected by the sort of strategic understanding to be sought on tier one and they, in turn, will determine whether the force improvements of tier two produce a stronger military coalition. The Grid can be seen as both network and knowledge base. Ensuring that the ones and zeros are passed correctly is no trivial challenge, but it is a concrete and measurable task on which progress is being made, thanks, in large part, to international standards both official and de facto. Ensuring that bits are understood correctly and responded to appropriately is not only technically but operationally challenging. After all, common situational awareness exists not only when people see the same symbols on the screen but also perceive their importance in ways that are either identical or can be bridged by the use of human intelligence. Issues of NATO Grid interoperability cannot be separated from issues of how U.S. and European forces bridge their differing tactical and strategic perceptions. The closer the United States and the allies are in defining the military missions and tasks their forces must be able to perform, the more likely a NATO Grid will work.

Construction. There are at least four alternative ways to build a NATO Grid:

An advantage of having each side build its own system and then merge them is that both sides start negotiating the terms of merger on the same foot. But that process still requires considerable trust, standards, and time, especially if the NATO Grid is to be as seamless as a national Grid. And requiring that systems integration be undertaken twice (i.e., on both sides of the Atlantic) or more often (if each large European nation goes its own way) will waste money.

Having the Europeans copy the U.S. architecture would give Europe an autonomous capability. Such a step would save initial outlays and make ultimate integration easier. The Europeans also would save some time and money required for systems integration. But waiting for the United States may reduce their capability in the interim and give the United States extraordinary leverage when the time comes to give out the integration techniques. Porting these techniques will be hard; there will be nontrivial differences between U.S. and European equipment and doctrine, the adjustments to accommodate them may get in the way of melding, and many of the techniques may be difficult to transfer without hefty and costly assistance from U.S. defense firms.

In the third option, the U.S. Grid would be supplemented by sensors, databanks, processors, and fire-control units owned and operated by the Europeans. European users, in turn, would access the U.S. Grid as U.S. users would. But similar access need not mean identical access. Connectivity does not have to override command arrangements (linking Dutch artillery to a NATO Grid built on a U.S. backbone does not, by itself, mean U.S. commanders can fire Dutch weapons). Nor does connectivity override need-to-know criteria that permit sensitive data to be compartmented. Granting such distinctions, the allies would forego the expense of doing their own systems integration,19 but they also would lack an autonomous capability for RMA-level operations.

In the fourth option, the United States and the Europeans would share the costs of building a NATO Grid from scratch and thereby become mutually interdependent. This option would waste the least effort and most strongly reinforce the alliance structure. Yet, the Europeans would be paying for capabilities they cannot use. The United States, for instance, may want the Grid to track a target in real time so as to convey updated coordinates to its next generation of long-range missiles, but the Europeans may have no such weapons. The United States must also think of other alliances (e.g., with Japan, Australia) and relationships (e.g., with Israel, Latin America). Designing the U.S. Grid around NATO assets may deny the U.S. a capability for autonomous RMA operations, a capability it needs far more than the Europeans do.

The best choice for the Atlantic alliance would combine a common architecture with some elements of a common backbone and service infrastructure--a hybrid of options two and three. The United States is 5 to 10 years ahead of the Europeans and cannot wait for them to catch up before building its own Grid. Yet, building a U.S. Grid without taking into account the likelihood of ultimate integration would fly in the face of alliance requirements. It could reduce the benefit of European force improvements and transatlantic convergence on global strategy and military requirements.

To make a NATO Grid work, the United States should be explicit about its architectural choices, be willing to share the key integration techniques with the Europeans, create enough flexibility so that the Grid can accommodate a range of doctrines and command arrangements, and use commercial standards20 to facilitate easy plug-and-play. To be blunt, the United States should be generous in order to pre-empt the Europeans developing a separate architecture. U.S. initiative and openness would leave Europeans a choice between sharing the U.S. backbone (and/or associated services) or developing their own architecture at great cost to themselves.

Agreement on how the Grid is to be built clears the way for working out features of the Grid that would best help both sides--features such as content, fair access, and security.

Content. Information is useless until put in a form compatible with users' weapons, doctrine, sophistication, expectations, and rules of engagement. If the United States military prefers to use dispersed forces, agile operations, and standoff precision strike to conduct its wars, the information it seeks would logically reflect that preference. Its C4ISR systems will scan large areas of terrain, sift the background to locate a few nuggets of actionable activity, sort them by priority, and send the information to strike units. Those who fight differently need different information. If the NATO Grid cannot support their requirements because it reflects U.S. inputs and designs, its value to allies will be modest, perhaps even negative.

Consider the act of identifying and locking onto a truck equipped with a Bushmaster-class machine gun. Current vehicle tracks are flashed to U.S. forces, which then strike from standoff range. Would these data suffice for allies? If their weapons are precise but lack range, allies would need to fire from close up; they would need to know where other enemy assets lay in order to operate from protected spots. If the allies' weapons are not so reliably precise, then allies would need trustworthy, real-time battle damage assessment for subsequent reengagement (preferably before the enemy shoots back). If allies use ground forces to smoke out adversaries, they would need the Grid to find the best way in and out quickly in order to shoot without being trapped in the chaos of small-arms exchange. By contrast, some forces shoot as soon as they have sufficient evidence that the target is more likely hostile than not, rather than wait for certainty beyond the reasonable shadow of a doubt that U.S. forces may require.

Differences in what information should be provided will depend on how vigorously Europeans pursue other RMA investments and doctrines. The more long-range precision weapons they acquire, the faster they adopt stealth and electronic warfare techniques, and the greater their capacity to deploy assets to distant theaters, the more they will be able to conduct standoff operations. These developments would make Europeans more inclined to use the same information that the United States would be generating for its own use. Nevertheless, the NATO Grid has to take inevitable differences in culture, politics, sensitivities, and weapon inventories into account when putting the right information into a common operational picture. The greater the agreement across the Atlantic on strategic and operational tiers, the simpler will be the process and the more seamless the results.

Fair Access. Similarly, a broad convergence between the United States and Europe on the strategic issues--tier one--will help each side trust a Grid to which each has contributed. But Europeans have to believe that a NATO Grid serves them fairly and is not simply a tool of the United States, a reflection in silicon of how the United States would go to war and would have its allies go to war as well.

What is "fair access?" Allies with a need to know ought to be able to upload and download information and tap into the Grid's knowledge and services on an equal basis. Such a Grid would span NATO without creating class distinctions--i.e., first-class Americans, second-class Europeans. If the Grid fits the needs of all users to a comparable extent, Europeans, no less than Americans, can be equally motivated to invest in and rely on it. The buy-in process also would help Europeans understand better how Americans think at the operational level of war (and vice versa). This, in and of itself, would make it easier to fight as a coalition

Some technological barriers to fair access ought to disappear over the next 10 years. Difficulties that foreigners have in even if reading and writing files in English can be overcome through translation programs. When coupled with voice recognition technology, they can facilitate video-teleconferencing or other collaboration over the Grid. Europeans have complained that they cannot keep up with Americans because they cannot afford the hardware (even if a realignment of Europe's defense priorities, as suggested in the last chapter, would give them more resources for investing in the Grid). Yet, a $1,000 desktop computer (early 1998 prices) can run all software not specifically written for such dedicated boxes as mainframes, supercomputers, and MILSPEC (military specification) hosts. Comparable mobile units are falling below $1,500. Digital mobile telephony is also getting less expensive.

Current field-level bandwidth constraints that favor some users over others can be overcome with sufficient resources. The 9,600 bits per second limitations of the Army's mobile units that constrain the Army's Force XXI architecture will soon pass or be correctly regarded as archaic. Used adroitly, commercial technology should be able to fill the gap. For example, Metricom's microcells can transmit 30 thousand bits per second to cards that slot into laptops, AT&T's cellular technology can get 128,000 bits per second to mousepad-sized antennae, and direct broadcast satellite technology can broadcast a billion bits per second to a receiver just over a foot wide. Elsewhere, bandwidth constraints can be eased by intelligent message processing, large local storage (digital video disks circa 2000 should be able to store 17 gigabytes, enough to hold a compressed full-color image of the former Yugoslavia accurate to a meter), and the substitution of standard symbols for some images. Fiber promises near-infinite capacity to all fixed facilities.

Fair access also entails ensuring that common operational pictures supported by the NATO Grid not reflect a U.S. bias. Even if Europe acquires RMA capabilities, Europeans may not look at combat exactly as Americans because of lingering differences in culture, equipment, and thus doctrine. Consider information that may be gathered about the safety of a village in a war zone. A possible arms build-up is reported nearby. Should the village be marked safe for planning relief operations? Those with rapid reaction forces may say yes ("our forces will intervene in time"). Those without may say no. Whose judgment goes into the common operational picture (or should both)? If a user wants backup for this judgment, where does (double-clicking on) the village icon lead? Further assume a local leader has closed a dirt runway serving the village. What areas--e.g., air operations, ground missions, and logistics planning--will be affected by this news? The answer will be influenced by how the knowledge bases which build the Grid's common operational picture are linked. If there are changes in the situation, which of them should be highlighted for human attention? The answer will affect how the common operational picture is presented.

Assuring Security. Melding two Grids guarantees security headaches. As a system grows larger, so does the number of entry points--negligent users, unknown ports, prisoners caught with laptops accessing the Grid, or even spies. Information warfare--specifically defending systems against malicious intrusion--presents another complicating factor. Adroit intruders can, in theory, read a system's information, feed a system false or misleading information, command devices attached to a system, or cripple a system's operations. If security officials in the United States and Europe do not trust their counterparts, the firewalls between the two grids may keep work from getting done. If both sides trust the other but pursue divergent security policies, hackers may be able to play off weaknesses in one system against those of another.21 If anything goes wrong, errors may be hard to trace, and responsibility difficult and contentious to assign.

That either the United States or Europe will be reluctant to extend its resources to the other without trustworthy security assurances is obvious. Fortunately, many categories of information previously considered too sensitive to share have been downgraded to NATO Secret so that allies can use them more routinely. Some content is being distributed more widely, even as sources and methods remain a deeper secret. Other data (e.g., the source of a signal) is distributed, while related data (e.g., the content of an intercepted signal) are closely held.

A NATO Grid needs compatible, commonly administered, and mutually trustworthy security features, such as encryption, authentication, filters, firewalls, reporting mechanisms, and anti-intrusion devices. Agreement on these features is hard enough to reach when allies have similar opinions on who can be trusted and how much; and, as the differing U.S. and Italian perceptions of how to deal with Somalia's warlords showed, such agreement cannot be taken for granted.

Creating Openness in a NATO Grid

For the United States to pursue the RMA track without paying attention to alliance needs up front may result in a magnificent U.S. System of Systems that must be laboriously disassembled and reassembled if it is to function for NATO. Or with the costs of conversion so large and deadlines for any given mission so close, the United States might forego the process entirely and thus lack the option of responding jointly with its allies even when common interests are threatened. Conversely, the earlier NATO needs are considered, the more the Europeans can feel that they have a stake in RMA success.

The components of a NATO System of Systems exist, but a fully integrated U.S. Grid, much less a NATO Grid, is in its infancy. The process of creating a NATO Grid will easily take a decade or more. In that time period, some interoperability problems (e.g., from divergent communications hardware, or language differences) may fade on their own. As new capabilities are developed, however, new incompatibilities may arise. Issues arising over system construction, content, access, and security are all potential barriers to free and easy access. All must be handled assiduously.

If the United States is serious about being able to broaden its Grid into something its allies can use, it must consider the ability to broaden the Grid as it weighs issues that will affect the Grid's architecture. It must ensure that the Grid is open, not only technically, but to accommodate allied operational preferences. Designers must be explicit about the key decisions, inform NATO allies of these, solicit their input, and give their opinions appropriate weight. Needless to add, allies will not always get their way. Accommodating their requirements may reduce capabilities, or add time and cost.

The Grid also must meet U.S. requirements in situations where NATO is not involved, and U.S. needs may not always match those of its allies. Nevertheless, a process that explicitly identifies the key design issues and lays out clear alternatives in their resolution at least makes clear what may be lost by building the Grid only to U.S. specifications, and thus makes more obvious what may be gained through sharing perspectives on global strategy, operational capabilities, and tactical doctrine.

Many features that make it easier to extend the Grid to the allies can also make it more robust, flexible, and adaptive:

The need for openness is more than a technocrat's mantra. Even if the United States and Europe succeed in forging a common strategic orientation, and Europeans equip significant forces with RMA equipment employed according to RMA doctrine, as earlier chapters argue, the actual ability of the United States and Europe to go to war together will be vitiated if U.S. forces are on the Grid and European forces are off the Grid or on an incompatible one. Openness, in turn, offers a Grid to which U.S. and European manufacturers could contribute on an equal basis, the cornerstone for a free transatlantic defense market, the subject of the next chapter.


Notes

15 Dividing the conflict area into zones for U.S. forces and zones for forces off the Grid would obviate the cross-fire problem--assuming actions took place within and not across zones. In the latter case, opportunities would be worse than lost, they would have been conceded beforehand.

16 See Larry Wentz, ed., Lessons from Bosnia: The IFOR Experience (Washington D.C.: Department of Defense Command and Control Research Program and National Defense University Press, 1998).

17 French peacekeeping forces in Bosnia, for instance, used real-time airborne video surveillance to monitor Mostar's polling places, which put them in a position to concentrate their forces against potential disturbances without needing to patrol polls directly. See the Defense Science Board, Improved Application of Intelligence to the Battlefield (Washington, D.C.: Office of the Secretary of Defense, 24 February 1997), 46.

18 Information services should be understood as access to data, network services (e.g., language translation), applications, and, if economics permits, data storage and processing facilities. Once established, their marginal costs are minimal, and they can be offered for little or no payment--if it is U.S. policy to do so. Communications services (e.g., uplinks to satellites) are likely to be charged for whenever the marginal cost of adding capacity is expensive. Yet, trends in switches and fiber optics suggest that it pays to build the fattest terrestrial pipes possible even if they cannot be filled with traffic anytime soon. The marginal cost of handling low-bandwidth applications such as voice traffic or E-mail is almost zero, and the U.S. military may be able to let the Europeans use the infrastructure for free and never notice the difference.

19 Europeans would have to write software to link their systems to a U.S. Grid, just as they would have to do for their own Grid. The more services the U.S. Grid offers to and requires from component systems the more complex such software would be (e.g., compared to what may be a less sophisticated European Grid).

20 Identifying the right commercial standards is easier said than done. Incompatible commercial standards often compete with each other. Some areas have standards under construction; yet others have none. Today's standards may not be tomorrow's. Because some standards are incomplete, choosing the right implementation dialect matters.

21 Assume System A is strict about keeping unauthorized users out but gives authorized users great scope to read and write system files and System B is more penetrable by unauthorized users but restricts all users' ability to read and write system files. If the two systems are linked too casually, a hacker can wreak havoc by penetrating System B to appear to be an authorized user, and then use this authority to invoke programs on System A that permit reading from and writing to sensitive system files.


Table of Contents  |  Chapter 5