‘Cyber ecosystem’ could mimic dynamics of the human immune system to isolate and contain threats robotically.
The federal government is strategizing to build a virtual community that would prompt computers worldwide to instantly, en mass, suppress cyberattacks, sometimes without humans at the keyboard.
The so-called cyber ecosystem would take “collective action” to galvanize cooperation among networks, external devices, and consumers, the Obama administration announced today.
“Computer systems, devices, applications and users will automatically work together in near real time to anticipate and prevent cyberattacks, automatically respond to attacks while continuing normal operations, evolve to address new threats, limit the spread of attacks across participating devices,” as well as share timely security information, a government research solicitation stated.
The Homeland Security Department and National Institute of Standards and Technology are seeking public input on the potential benefits and challenges of the approach before forging ahead with construction.
“This information will help DHS and NIST develop future cyber ecosystem security capabilities and an implementation strategy that will strengthen the security of critical infrastructures, federal information systems and the private sector,” the bulletin stated.
This initiative would bring to life a 2011 thought paper by Philip Reitinger, former deputy undersecretary for Homeland Security's National Protection and Programs Directorate, who has since joined Sony Corp. The entertainment giant was victimized by hacktivists who filched personal information on almost 100 million users of Sony’s PlayStation Network and Qriocity music service.
Cyber ecosystem dynamics could possibly mimic the way “the human body reacts to an infection,” federal officials said. Following the analogy, the system, for example, might be engineered to react locally to the immediate threat, or point of injury, as well as globally to contain damage.
Such an ecosystem would require worldwide buy-in from industry, governments, universities and consumers, officials acknowledged. They did not suggest a timeline for launch.
Monday’s proposal poses several questions to guide feedback from experts. It asks, for instance, how cloud computing -- relying on remote machines to process information -- could play a role in the collective action.
The operation likely would involve some degree of machine learning so that computers could respond in real time and adapt to changing network environments.
The solicitation partly aims to find out if the field of artificial intelligence is mature enough to support the concept outlined. One question posed, for example, is this: “Can the cyber ecosystem be self-learning and automatically apply the learning to an attack without imposing unacceptable consequences to the ecosystem?”
Federal officials also are examining the drawbacks of such a system, including unintended consequences and “privacy and civil liberties concerns associated with automated collective action, including information sharing between cyber devices, interoperability of systems, behavior monitoring and authentication.”
Monday’s notice also asked whether a public-private partnership should manage the program at the outset.
NIST currently is organizing a public-private venture aimed at developing a login network akin to the credit card payment system that would enable Internet users to access distinct websites and apps without reentering personal information or creating new passwords. In August, the program, called the National Strategy for Trusted Identities in Cyberspace, convened a steering group headed by officials from the private sector, including PayPal, Citigroup and Aetna.