Anatomy of Network-Centric Warfare
Ubiquitous connectivity may have hidden consequences.
As momentum grows for network centricity in military operations, architects of the plans may find themselves closely examining sciences such as sociology or biology to preview where network-centric activity can lead. When command, control, communications, computers and intelligence systems become more highly networked, the need for sophistication in the products and platforms that sit at the edges diminishes. In some cases, too much capability at the edge may actually inhibit self-organizing behavior and negatively impact the mission of the networked whole.
The term network-centric is an oxymoron. Networks themselves are inherently centerless—it is all about the edges. The promise of networks is in the dynamic interactions of the nodes at the edges. When these nodes become highly connected, interesting things start to happen.
The fundamental principle that underlies today’s thinking about network-centric warfare is Metcalfe’s Law, named after Robert Metcalfe, the inventor of Ethernet and founder of 3Com Corporation. It was a phenomenon first observed in commercial communications as Metcalfe sought to address the problem of creating larger networks out of many smaller ones. The law contends that the power of a network increases with the square of the number of nodes connected to the network. Network-centric warfare advocates build on this law by asserting that maximizing the number of nodes increases the chances of realizing the promise of networks through ubiquitous connectivity and interoperability.
Pursuit of ubiquitous connectivity may have hidden consequences, however. Carried to its logical conclusion, such connectivity may give rise to an environment that becomes increasingly “biological” over time and begins to exhibit behavior similar to that of living systems. However, living pathways represent everything that the traditional business of defense is not. The implications of biological constructs in defense are both surprising and scary.
Albert-László Barabási and Eric Bonabeau outlined recent discoveries regarding highly networked systems in an article in Scientific American magazine. Chief among these discoveries is the existence of power laws that underlie networks such as the World Wide Web or the relationships between people in society. The power law behind these networks means that they are inherently scale-free—there is no typical node and no peak in the distribution curve. In networks of this type, the vast majority of nodes have a small number of links but are juxtaposed with a tiny number of nodes that have an extraordinarily large number of connections. The latter are known as hubs.
Hubs are evident in everyday life. In society, a few people operate in a huge circle of acquaintances amidst a world where only a limited number of relationships are nurtured. This also is evident in the disproportional popularity of certain Web sites such as Amazon or Yahoo when compared with nearly every other Web page. Interestingly, these hubs also are present in the command and control (C2) hierarchy and constructs that are currently employed in the business of defense.
Complex networks operate through the dynamic interaction of nodes and hubs that resemble an unfolding drama more than an engineered system. Perhaps the most conspicuous characteristic of these networks is that they continuously grow. In the push toward connectivity growth in network-centric warfare, the U.S. military may be unwittingly marching down a path that conforms to the logic of scale-free networks. The question is whether this is good or bad.
Barabási and Bonabeau ominously warn that while scale-free networks are quite survivable against random failure, they also are highly vulnerable to a focused attack because of the critical role of hubs in the network. Their research has shown that up to 80 percent of the nodes in a scale-free network can be randomly removed without compromising the functionality of the networked whole. However, a coordinated attack that disables 5 percent to 15 percent of the hubs simultaneously is all that is needed to mortally wound such a system.
Ubiquitous connectivity of a scale-free variety encourages a dangerous centrality of interaction focused on C2 hubs. Dynamic tension exists between hierarchical C2 models and the push for maximizing the number of nodes. While maintaining a hierarchical C2 model, maximizing the number of nodes is a risky proposition because it amplifies the network’s vulnerabilities. However, if maximizing the number of nodes is the overarching imperative, then new models of C2 may emerge. These frameworks may be more efficacious in monitoring, pre-empting and combating threats of an asymmetric and fluid nature.
But these new C2 models will likely have to dispense with traditional notions of control and efficiency to accentuate flexibility and adaptability. When networked information flow is nonexistent, hierarchy rules. When it is abundant and dynamic, the traditional C2 pyramid gets flipped upside down and the edges become primary. Command intent gradually diminishes in importance in comparison to emergent intent that arises spontaneously through the interaction of autonomous agents at the edge.
A scale-free philosophy also implies nodal expendability—by design. Small failures are bred into the system so that fatal breakdowns are avoided. In today’s ethos of casualty avoidance, how does this notion get initiated in battle planning wherein up to 80 percent of the nodes are potentially expendable without mission degradation? When such odds are possible, creation of operational concepts that leverage such odds in favor of overall mission accomplishment will naturally emerge to the benefit of hubs and to the potential detriment of most nodes. From an operational perspective, the needs of the few outweigh the needs of the many if the overarching good of all is to be achieved. The likely consequence of such a construct is the increasing substitution of machines for humans at the nodes. Machines are expendable. Humans are not. This trade-off is already starting to occur in earnest.
The interaction of highly networked systems is not governed by mechanisms that traditional systems engineering techniques can necessarily discern or manage. Their nonlinear characteristics can frustrate the drive for precision, optimization and predictability—the hallmarks of military products, platforms and systems. The messy redundancy of networks has a recursive nature that manifests itself in curious ways.
Author Kevin Kelly, co-founder of Wired magazine, talks about simulations that illustrate the dynamics of networked interaction. They reveal a fragile balance between nodal intelligence and range of capability versus networked outcomes. In one simulation, the autonomous, self-organizing actions of thousands at a computer conference enable them to fly and land an airplane effectively without any direction from a coordinator. The key to mission accomplishment, however, was the limited range of behavior accorded to each participant in the form of a joystick that enabled only up/down and left/right movement.
A follow-on simulation that took place several years later involved a submarine in search of targets. This time, the results were not initially successful. The primary difference was that participants were accorded a much higher level of sophistication and control. The increased range of behavior resulted in chaos as individuals and groups of individuals nullified each other’s actions in the exercise of their autonomy. Success was achieved only after the simulation director stepped up to the microphone and provided some explicit directions to the entire group, which then enabled swarm behavior to be unleashed effectively. That direction was the visible hand of command and control.
The lessons of these simulations are interesting. For example, a highly networked environment provides the means for less capable legacy assets to be reborn with new purpose and meaning. Kelly has quipped that “dumb parts properly connected can yield smart results.” Under certain conditions, ever increasing quantities of “dumber” nodes linked to the network are likely to enhance mission effectiveness as much if not more than adding more sophisticated ones.
However, achieving synergy among intelligence, quantity and control is a tricky proposition. In a network-centric warfare environment, C2 becomes more of a coaxing and tweaking phenomenon. C2 will gradually look more like the Federal Reserve chairman’s attempts to influence the economy through soft levers like interest rate adjustments rather than through concrete orders and actions.
Biological constructs, much like world economies, however, sometimes give rise to wild-card factors that are not anticipated. The emergent characteristics are often rooted in the profusion of organic systems exerting influence laterally. These systems operate in a peering arrangement involving hundreds of thousands and even millions or billions of nodes. Where is the cross-over point in which unexpected effects begin to appear as the world of big numbers and horizontal causation unfolds? As military operations begin to make greater use of swarm systems in the form of networked microsensors, robots and unmanned vehicles in concert with humans, an understanding of thresholds separating discrete control versus unpredictability is critical.
The inner workings of biological systems are oftentimes opaque to linear modes of discovery. Stephen J. Gould, the late paleontologist from Harvard University, once posed the question, “What good is 5 percent of an eye?” A traditional systems engineering approach to problem solving might result in the following response: “Well, seeing with 5 percent vision is certainly better than having no eyesight at all just like having 10 percent vision is better than having only 5 percent.” But there is a fundamental flaw in this logic. Phillip E. Johnson, professor of law at the University of California at Berkeley points out that 5 percent of an eye does not equal 5 percent vision. The eye is just a portal into an extraordinarily complex network of systems known as vision. This kind of twist is commonplace in biological systems.
The Internet is another example of such a twist. Born as a defense project to facilitate communications in the event of nuclear attack, it has transformed into a successful commercial invention. This outcome was not intended nor was it foreseen, and it was a direct result of the biology of highly networked constructs.
Because the wildness of living systems lies in wait and is only evidenced by the overall effects of the networked whole, the focus of modeling and simulation in defense will change over time. Increasing effort will be placed on distributed modeling and simulation of abstract elements at play. Of primary concern will be the unintended consequences that arise through the self-looping interaction of zillions of nodes. Future defense simulations may actually look, feel and operate more like the SimCity line of computer games from Electronic Arts than mechanisms currently in use. This transformation will only intensify as the world becomes engulfed in a sea of microsensors and telemetric devices marching inevitably toward universal connectivity.
A robustly networked environment also has implications for human capital investment. Demand for traditional defense skill sets like electrical engineers and computer scientists will not disappear. However, standing next to them, like weeds popping up in a lush garden, will be workers of a different sort. These agitators against the accepted order will be people schooled in disciplines such as sociology, zoology, linguistics and the cognitive sciences. These individuals can better understand what happens when organic characteristics start to become expressed in aerospace terms. As military operations become truly network-centric, these arcane disciplines will form the basis of a new tail that will eventually wag the defense dog.
It is clear that the migration toward network centricity carries hidden implications. By maximizing connectivity and interoperability to realize the promise of networks, biological characteristics may eventually emerge. When they do, the defense industry as a whole may transform in ways not yet imagined.
Clement C. Chen is vice president, strategic analysis and development, for Lockheed Martin Corporation, Bethesda, Maryland.