S Lab

[Sibin Mohan]

Systems Security Research Group at GWU and University of Illinois

Integrating Security as a First Class Principle in Real-Time Systems

Team Members: Sibin Mohan, Monowar Hasan

Collaborators: Rakesh Bobba, Rodolfo Pellizzoni

Embedded devices with real-time properties (e.g., strict timing and safety requirements) require interaction between cyber and physical worlds. These devices are used to monitor and control physical systems and processes in many domains, e.g., critical infrastructures like the electric grid, oil and gas pipelines, water systems, EDS control networks, process control systems in industrial plants, unmanned aerial vehicles (UAVs), submarines, electric vehicles to name just a few.

*Image courtesy: Krypt3ia

Given the drive towards remote monitoring and control, these devices are being increasingly interconnected, often via the Internet. Since many of these systems have to meet stringent safety and timing requirements, any problems that deter from the normal operation of such systems could result in damage to the system, the environment or pose a threat to human safety. Similar problems can occur if such systems are the target of malicious security attacks – a very valid concern since breaches are not uncommon in critical applications, especially considering the recent spate of IoT-centric attacks.

Enabling security in real-time systems (RTS) is often more challenging than generic enterprise systems due to additional real-time requirements (such as limited energy, safety and timing constraints). Stringent timing constraints in RTS introduce complexities – the strict deadlines for the real-time tasks may not allow for frequent execution of security mechanisms. For example, how often and how long should a monitoring and detection task execution to be effective but not interfere with real-time control or other safety-critical tasks? While this issue potentially be addressed for newer systems at design time, this is especially challenging for retrofitting legacy/existing systems where the real-time/control tasks are already in place and perhaps cannot be modified.

Illustration of an adaptive security integration mechanism: by default security tasks (say a given set of intrusion detection policies) execute with lowest priority (e.g., as a background process) – once any anomalous behavior is detected, the tasks get higher priority and perform more fined-grained system-wide checking.

The major focus of this project is to address the following:

how do we integrate and then characterize the effects of security in legacy time/safety critical systems (those designed using both single and multicore chips)?


For legacy/existing systems

Future systems