Trust Measurement Methods in Organic Computing Systems by Direct Observation

Abstract

Trust is an important factor for autonomous distributed systems. Organic Computing features self-organizing techniques to manage complex distributed systems. This paper introduces two methods to measure and calculate trust values with respect to the reliability of other nodes by direct observation introducing minimum overhead. We demonstrate this approach using a simulator and show that both algorithms converge towards the fixed real trust value and stay within close bounds of it. Evaluations also show the amount of measurements needed to calculate a meaningful trust value as well as the importance of the nodes' own reliability value when measuring another node. Additionally, we introduce an enhancement of one method to calculate a meaningful trust even by an unreliable observer.