Security-Cognizant Real-Time Scheduling

Abstract

The often very limited computational capacity available upon resource-constrained safety-critical embedded systems (such as unmanned aerial vehicles) must be carefully managed in order that they may provide the enhanced functionalities expected of them. Such systems are increasingly becoming a target of hacker attacks; hence there is an increased recognition of the need for real-time resource-allocation techniques that are resistant to such attacks. To meet this need we (i) propose a natural security-cognizant extension of the sporadic task model that is currently widely used in real-time computing; and (ii) develop a scheduling algorithm and associated schedulability test for this task model that guarantees correctness of both timing and security properties.