By Jay Landers
As critical infrastructure becomes more automated and interconnected, the potential threats against it likewise increase in terms of their unpredictability and severity. To protect infrastructure from bad actors who would attempt to exploit cybersecurity vulnerabilities, the engineering research and development community must take a more proactive approach to address such weaknesses, rather than simply developing reactive measures to protect against known threats.
That is the conclusion of a report recently released by the Engineering Research Visioning Alliance. The ERVA is a five-year initiative, funded by the National Science Foundation, that seeks to promote research within the engineering community that addresses key national, global, and societal needs.
‘Visioning event’ findings
In mid-February, the ERVA released a report, titled Engineering R&D Solutions for Unhackable Infrastructure. The report summarizes the findings from a “visioning event” conducted by the ERVA on Aug. 10-11, 2022, at the Massachusetts Institute of Technology.
For the event, the ERVA convened nearly three dozen researchers and experts from academia, government, and industry to “identify engineering research priorities leading to significant advancement in the security and resilience of tomorrow’s cyber-physical infrastructures,” according to the report. “The visioning exercise addresses longstanding trends in national and global scale cyberattacks on our increasingly complex and cyber-enabled infrastructures.”
“Anticipation and preparation for security threats to physical and virtual spaces has emerged as a critical national and international priority,” said Saurabh Amin, Ph.D., an associate professor of civil and environmental engineering at MIT and a co-chair of the task force that led the event, in a Feb. 16 news release.
“The research priorities described in this report provide the necessary steps we must take to mitigate the economic and security threats inherent in cybercrime and secure both physical and virtual spaces,” Amin said.
With cybercrime increasingly on the rise, the visioning event participants sought to address the question of how to shift infrastructure security and resilience research from being “merely reactive” to being “anticipatory,” according to the report.
“Visioning event participants asked what tomorrow’s ‘unhackable infrastructure’ could look like,” according to the report.
The participants identified five key areas as priorities for engineering research:
- Human-technology interface considerations.
- Measuring and verifying security.
- Future approaches to autonomous security.
- Components and resilience in interdependent infrastructures.
- Architecting trustworthy systems.
The human role
For the first priority, the experts noted that humans merit more attention from researchers, given the key roles they play in the realm of security.
“Humans are simultaneously the weakest links in security yet the greatest opportunity to protect cyber-physical infrastructure,” according to the report. This is because they will design, deploy, operate, manage, and evaluate future cyber-physical systems, but they will also “be the adversaries who engineer hacks, intrusions, abuses, and attacks against tomorrow’s infrastructures.”
The concept of cyber-physical infrastructure systems needs to be recast as cyber-physical-human infrastructure systems, with greater emphasis placed on human incentives, the report notes.
“Research is needed to better understand the dynamics of how humans use and abuse technology, and how human incentives can more systematically be anticipated and guided within a secure and resilient design,” according to the report.
Balancing security and usability is critical, so that security improvements do not limit functionality or increase inconvenience or cost, the participants noted. Another need involves enabling operators to better identify potential adversarial activities and develop defenses.
To help with this goal, more research is needed “to explore and enable the use of augmented and virtual reality (AR/VR) specifically for security systems,” the report notes.
Measuring and verifying security
When it comes to the second research priority, more research is needed to address the inherent challenges of measuring, evaluating, and verifying security in cyber-physical-human systems, the report notes.
Because infrastructure systems “will face a multitude of attacks and failures,” they will need “fully automatic mechanisms that can maintain key functionality (resilience) while rapidly recovering to a desirable operational state (recovery),” the report notes.
Therefore, “a potential avenue for future research is how to integrate predictive information about attacks (and their potential impacts) into autonomous decision-making agents that can trigger additional information-gathering tasks (e.g., through activating sensors) and defense mechanisms to reduce time-to-recovery and prevent system-wide failures,” according to the report.
For the third research priority, concerning autonomous security, the event participants noted that such systems are “sorely needed,” the report noted.
In large part this is because the scale and complexity of cyber-controlled infrastructure and the accompanying threats can exceed the capability of humans to monitor.
When it comes to security, autonomous systems “could provide real-time defenses against a newly discovered threat or modify moving target algorithms automatically to counter a pattern in adversarial behavior,” according to the report.
Ultimately, such systems need to become better at distinguishing between normal and hostile activity.
Concern for components
Along with being spread out physically, critical infrastructure often comprises different, complex components that are connected by communication networks.
Software bugs and hardware malfunctions can expose such components to adversarial attacks. The fourth research priority, new approaches to resilience in interdependent infrastructures, aims to address these vulnerable access points that could be exploited by hackers.
More work is needed to ensure the integrity of the various components that make up infrastructure systems. “An important opportunity for future research is to develop a design approach that maintains system-level properties of safety and security after integration of modular components,” the report states.
A trustworthy infrastructure system is one that “functions exactly as expected within its context of use, even when unexpected inputs or failures occur,” according to the report.
However, knowing when a system is functioning correctly can be a difficult task, one made all the more challenging by the complexity of such systems.
“While the space of correct system actions may be large, the space of possible incorrect actions is much larger and seemingly intractable,” the report notes. “Yet this is precisely the space adversaries occupy as they look for unexpected vulnerabilities and abuses.”
The fifth research priority, architecting trustworthy systems, calls for research to help “design and build trustworthy systems that are resilient to innumerable malicious inputs and crafted modes of system abuse intended to derail system function and/or hijack system control,” according to the report.
“The challenge is how researchers can better address the problem of specifying system constraints to manage the vast space of inputs and manipulations that are simply out of spec within (cyber-physical-human system) infrastructures.”
The research priorities called for in the report require attention if the infrastructure of the future is to continue functioning properly in the face of increased cyberattacks, said David Ott, Ph.D., a senior staff researcher at the computer security firm VMware Research Group and Amin’s co-chair, in the Feb. 16 news release.
“Critical infrastructures of tomorrow will face an enormous array of adversarial threats,” Ott said.
“Our goal is to identify research and development avenues in infrastructure security that will raise the bar on robustness and resilience,” he said. “New approaches in autonomous security and trustworthy architectures could be game changing.”
This article first appeared in Civil Engineering Online.