Countering Active Attacks on RAFT-based IoT Blockchain Networks

TL;DR

This paper analyzes active attacks on RAFT-based IoT blockchain networks, deriving coverage probabilities under jamming and proposing a fingerprint-based method for impersonation detection, with simulation results validating the approaches.

Contribution

It introduces a stochastic geometry-based analysis of jamming effects and a novel fingerprinting method for impersonation detection in IoT blockchain networks.

Findings

Coverage probability increases as jammers move away.

Error probabilities decrease with better link quality.

Proposed methods improve security against active attacks.

Abstract

This paper considers an Internet of Thing (IoT) blockchain network consisting of a leader node and various follower nodes which together implement the RAFT consensus protocol to verify a blockchain transaction, as requested by a blockchain client. Further, two kinds of active attacks, i.e., jamming and impersonation, are considered on the IoT blockchain network due to the presence of multiple {\it active} malicious nodes in the close vicinity. When the IoT network is under the jamming attack, we utilize the stochastic geometry tool to derive the closed-form expressions for the coverage probabilities for both uplink and downlink IoT transmissions. On the other hand, when the IoT network is under the impersonation attack, we propose a novel method that enables a receive IoT node to exploit the pathloss of a transmit IoT node as its fingerprint to implement a binary hypothesis test for transmit node identification. To this end, we also provide the closed-form expressions for the probabilities of false alarm, missed detection and miss-classification. Finally, we present detailed simulation results that indicate the following: i) the coverage probability improves as the jammers' locations move away from the IoT network, ii) the three error probabilities decrease as a function of the link quality.