Evaluation of Trust Metrics in an Artificial Hormone System

Organic Computing (OC) is a well-known research field aiming to build dependable embedded systems. OC systems often employ self-X properties such as self-configuration, self-healing, etc. The artificial hormone system (AHS) is a decentralized OC middleware, designed to autonomously distribute tasks on a heterogeneous processor network. The decentralized task distribution is controlled through short messages called hormones. The ever-growing array of application scenarios, as well as the large number of heterogeneous connected devices, enables a broad array of attack vectors, both inside and outside of the AHS’s control. Security mechanisms for the AHS have long become a necessity.In this paper, we propose the application of inter-processor trust, to address this issue on the AHS level. The presented trust concept is explained in detail, demonstrated in different attack scenarios and evaluated concerning its memory usage, communication overhead and efficiency.