Introduction
As the Internet of Things (IoT) continues to expand, blockchain technology is frequently proposed as a decentralized framework for managing trust and data integrity. However, most existing research assumes a stable, wired backbone—a luxury not afforded to many IoT deployments. In real-world wireless environments, blockchain nodes are highly susceptible to physical-layer attacks, most notably malicious radio jamming. This paper presents the first analytical framework to evaluate the security of RAFT-based wireless blockchain networks under such conditions.
System Model and Methodology
The study focuses on the RAFT consensus algorithm, favored in private and consortium blockchains for its high throughput and low latency. RAFT operates through a centralized "Leader" node that coordinates with "Follower" nodes. For a transaction to be successfully committed, a majority (over 50%) of followers must receive a proposal from the leader and return a confirmation.
To model this, the authors utilize a Poisson Point Process (PPP) to describe the spatial distribution of follower nodes in a 2D plane, with the Leader at the center. The communication is divided into two critical phases:
- Downlink (DL): The Leader broadcasts to Followers.
- Uplink (UL): Followers reply to the Leader.
Success is determined by the Signal-to-Interference-plus-Noise Ratio (SINR). If the SINR at a receiver falls below a specific threshold ($\beta$), the communication link fails. The paper's primary contribution, Theorem 1, provides a piecewise analytical expression for the probability of consensus success based on jammer location, power, and network SINR requirements.
The Geometry of Consensus Failure
A key insight of this research is that jamming does not affect the network uniformly; rather, it creates distinct geometric "zones of failure." In the Downlink phase, the jammer creates a circular "dead zone" around itself where followers are unable to receive the leader's broadcast. Conversely, in the Uplink phase, the jammer creates a "protection radius" around the leader; followers located outside this radius are too distant for their signals to be heard by the leader over the jammer's noise.
Consensus only succeeds if the combined "viable area"—where both Uplink and Downlink communications are functional—covers more than 50% of the total network area. The research demonstrates that the network is extremely sensitive to the SINR threshold. For instance, a shift from -10 dB to -6 dB can cause the success probability to drop from near-certainty to zero, effectively stalling the blockchain.
Results and Analysis
The analytical model was validated using Monte Carlo simulations in a 100-meter radius environment with 1,000 nodes. The results showed:
- Model Accuracy: Analytical curves matched simulation data almost perfectly, confirming the framework's utility for network planning.
- Jammer Proximity: Jammers are most destructive when positioned close to the Leader, as they can simultaneously disrupt the Uplink responses from the entire network.
- Path Loss Resilience: Interestingly, higher path loss exponents (representing obstacle-heavy environments) can make the network more resilient to distant jammers, as the jamming signal attenuates more rapidly than the desired signal over short distances.
Conclusion and Future Work
This framework provides a robust tool for assessing the physical-layer security of wireless blockchains. While currently limited to a single stationary jammer, future work aims to extend this model to multi-jammer scenarios and mobile topologies. The authors also suggest that predictive machine learning could enable networks to dynamically adjust parameters, such as transmission power or leader selection, to mitigate jamming attacks in real-time.
Summary updated