PROACTIVE INFORMATION SECURITY MECHANISM IN REAL-TIME OPERATING SYSTEMS USING A HYBRID WATCHDOG

Abstract

The paper introduces an advanced method for the preventive restart of real-time operating system components that integrates a simplified Markov reliability model with a hybrid dual watchdog timer architecture. This approach enhances cyber resilience, fault tolerance, and service continuity in the presence of both internal malfunctions and deliberate external influences such as denial-of-service attacks, logic bombs, and fault injection techniques. The use of compact Markov chains makes it possible to rapidly evaluate the probability of approaching a critical degradation state without imposing a substantial computational burden on the system. As a result, developers can ensure timely localization of faulty processes and carry out selective restarts of individual tasks, hardware drivers, or kernel modules instead of triggering global reboots.

A key contribution of the method is the integration of predictive and hardware-based protection layers into a unified multilayer mechanism. The predictive watchdog timer, operating at the stochastic modeling level, forecasts potential service disruptions based on transitional probabilities and observed runtime behavior. In parallel, the hardware watchdog component serves as a failsafe instrument, guaranteeing system recovery in situations of deep software blocking, escalated attack scenarios, or inconsistent module states. This dual structure reduces the frequency of full system restarts, limits downtime, and extends the operational life of critical components in resource-constrained environments.

Unlike approaches that rely on complex machine learning algorithms or burdensome formal verification techniques, the proposed solution is based on minimalist stochastic schemes that can be easily adapted to different platforms, architectures, and threat models. Experimental evaluation confirmed a more than threefold reduction in recovery time, a substantial decrease in reliance on global reinitialization procedures, and a noticeable improvement in information security indicators. The method thus contributes to increased reliability, uninterrupted control, and secure runtime performance in embedded and cyber-physical systems used in industrial automation, autonomous robotics, automotive electronics, and medical technologies, where failure tolerance and continued operation are paramount.