SHIRT (Self Healing Intelligent Real Time) Scheduling for Secure Embedded Task Processing

摘要

Scheduling in FPGAs are increasingly being employed in modern real-time embedded systems, which often impose strict timeliness constraints. Deploying such a schedule requires reconfiguration at various time instants with soft IPs procured from various vendors. However, performance degradation may be associated with a compromised soft IP from an untrustworthy vendor. The present work aims at tackling such a problem with a self aware approach. The security module checks the course of the proceedings at each intermittent point of a schedule and on detecting a malicious environment, heals the scenario and takes precautions to prevent similar malfunctions in future, without hampering the pre-determined schedule. Experimental validation shows effectiveness of our proposed methodology.