RT-SPDM: Real-time security, privacy and dependability management of heterogeneous systems

Abstract

The need to manage embedded systems, brought forward by the wider adoption of pervasive computing, is particularly vital in the context of secure and safety-critical applications. This work presents RT-SPDM, a framework for the real-time management of devices populating ambient environments. The proposed framework utilizes a formally validated approach to reason the composability of heterogeneous embedded systems, evaluate their current security, privacy and dependability levels based on pre-defined metrics, and manage them in real-time. An implementation of Event Calculus is used in the Jess rule engine in order to model the ambient environment context and the rule-based management procedure. The reasoning process is modeled as an agent’s behavior and applied on an epistemic multi-agent reasoner for ambient intelligence applications. Agents monitor distinct embedded systems and are deployed as OSGi bundles to enhance the real-time management of embedded devices. A Service Oriented Architecture is adopted, through the use of the Devices Profile for Web Services standard, in order to provide seamless interaction between the framework’s entities, which exchange well-formed information, determined by the OASIS CAP standard. Proof-of-concept implementations of all entities are developed, also investigating user-friendly GUIs for both the front-end and back-end of the framework. A preliminary performance evaluation on typical embedded devices confirms the viability of the proposed approach.