Wireless cyber warfare: Why mobile networks pose great risk

In cyber warfare, the network is the battlefield. While all networks are vulnerable to attack, mobile wireless networks are the most unprotected because their strengths and benefits — agility, adaptability, node autonomy and self-organization — also make them harder to defend against radio frequency (RF) distortion and malicious packet-level disruption and intrusion.

Up to now, wireless is the most neglected network security domain in terms of spending, in both military and enterprise spaces. Yet wireless networks, especially mobile networks, are the most critical component of tactical communication infrastructure and most challenging to defend against cyberattacks.

Whether relying on an impromptu network of smart phones or emerging technologies like the Joint Tactical Radio System (JTRS), the benefits of mobile ad hoc network architecture make it hard to distinguish between malicious packet loss and loss from environmental effects such as RF interference and rugged terrain. Mobile ad hoc networks are particularly at risk because they require node autonomy and self-organization to ensure robustness. Network-wise attackers can capitalize on the numerous network algorithms and protocols, such as ad hoc routing, which assume that all nodes are cooperating with the same goal in mind.

Even passive eavesdropping can be used to reveal the location of other network nodes, and the traffic pattern can be used to deduce other strategic information. If a wireless device is physically captured or hijacked, it risks revealing location information and packet contents while the rest of the network remains unaware.

The most secure mobile wireless networks must therefore operate in a trust-but-verify mode with every other peer node and protect information, including routing information, at every layer of the network stack.

The most neglected holes, however, continue to lie in the higher layers of the stack. Unlike the previous generation of communication technologies that used intermediate nodes primarily as relays, the Future Force communication technologies rely on routing as an integral part of the communication architecture. Thus, a new set of vulnerabilities are exposed to exploitation by a competent adversary using techniques like worm-holes. An adversary need only shut down a network via distributed denial-of-service attack and listen to it coming back up through “routing holes” to gather critical information about its structure. Many of the cyber warfare scenarios envisioned revolve around a single compromised node with the goal of preventing the red forces from infiltrating the tactical edge. Once they’re in, the entire large-scale network and all applications are exploitable.

To address this, focus is needed on the scalability issues of a wireless cyberattack.

In 2008, the United States launched the Comprehensive National Cyber Initiative, leading to the creation of the National Cyber Range (NCR) to be the site of comprehensive research on cyber systems. That research will involve modeling wireless technologies used in various radio systems, such as the Single Channel Ground and Airborne Radio System, TacSat, and JTRS radios, and other devices (including computers and smart phones) connecting to logical points across the test networks, including nongovernmental networks such as cyber cafes and metropolitan wireless infrastructure. To this end, the NCR must have the ability to emulate large computer networks that incorporate wireless communications as first class citizens over networks configured in operationally relevant scenarios, including net-centric applications and middleware and factoring in how people operate and use these tools.

It is not yet known if future on-the-move communication networks can be made secure enough. Given what’s at stake to meet cyber defense and cyberattack objectives, it’s critical that network designs, applications and users are stressed in an ultra-high fidelity and complex virtual space environment that accurately mimics the environment they will need to survive in.

Software Virtual Networks (SVNs) which the military is beginning to adopt, make it possible to represent the communication infrastructure at such high levels of fidelity that applications running on it –– such as a mix of third-party streaming video, voice over IP, e-mail, chat, video Web conferencing, video teleconferencing –– can be deployed unmodified on top of large emulated networks of both legacy and future communication devices.

The importance of incorporating real users and their exact applications is articulated in many recent observations by the professionals wrestling with the cyber challenge: “What keeps me up at night are poor browser and SQL database configurations,” said Ray Letteer, who heads the Marine Corps’ IA Division in the Office of the Director, C4/DON Deputy Chief Information Officer. “My blue teams that do operational tests keep finding issues." We need to mimic these precise operational environments in a virtual context in order to close the holes.

“Just as corporate users are vulnerable when they connect to enterprise networks using home Wi-Fi connections, soldiers are at their most vulnerable when they use wireless communications in crowded urban environments,” said John Morrison, chief of Lockheed Martin’s Wireless Cyber Security Lab.

Letteer acknowledges that urban battle settings are difficult to recreate in a laboratory environment; we suggest this can be effectively addressed through employing SVNs to directly support the creation of alternative test environments.

Further, to fully understand cybersecurity, real people need training to learn the tactics, techniques and procedures of cyberwarfare using the actual mix of applications and capabilities that they will be equipped with in the field.

What is needed to address cyber threats in the wireless security domain are solutions to respond to attacks at the higher layers. Just as importantly, we need a valid way to assess the cost/benefit of these solutions in an operationally relevant context.