A Privacy-Preserving Markov Reputation Framework with Local Identity Binding for OIOT Networks

Abstract

Focusing on store-carry-forward communication and trust-aware routing when Internet of things is sporadically connected, opportunistic Internet of Things (OIOT) networks can enable identity-reputation correlation, behavioural profiling as well as reputation fraud. This paper introduces an AES-GCM-encrypted reputation storage/private key, HMAC-SHA256 integrity checks, and PID-associated authenticated data/privacy-preserving Markov reputation model that involves node-local pseudonymous identity binding, no global reputation exchange, and HMAC-SHA256 integrity checks. One of the considered insider forwarding threats is blackhole behaviour. In the Opportunistic Network Environment simulator, the framework was tested on 30 paired runs of the four scenarios: Baseline Epidemic, Blackhole, Markov, and Privacy-Markov. Privacy-Markov achieved 0.28 delivery probability, 2.17 overhead ratio, 6,297s latency, and 1.267 hops, with low variability observed across the paired runs. No identity-exposure, reputation-sharing, integrity-verification-failure, or AAD-authentication-failure events were recorded under the evaluated conditions; however, cryptographic execution time and energy consumption were not modelled.