A Jamming-Resistant MAC Protocol for Multi-Hop Wireless Networks

This paper presents a simple local medium access control protocol, called \textsc{Jade}, for multi-hop wireless networks with a single channel that is provably robust against adaptive adversarial jamming. The wireless network is modeled as a unit disk graph on a set of nodes distributed arbitrarily in the plane. In addition to these nodes, there are adversarial jammers that know the protocol and its entire history and that are allowed to jam the wireless channel at any node for an arbitrary $(1-\epsilon)$-fraction of the time steps, where $0<\epsilon<1$ is an arbitrary constant. We assume that the nodes cannot distinguish between jammed transmissions and collisions of regular messages. Nevertheless, we show that \textsc{Jade} achieves an asymptotically optimal throughput if there is a sufficiently dense distribution of nodes.

💡 Research Summary

The paper tackles the problem of robust medium‑access control in multi‑hop wireless networks that are subject to a powerful adaptive jammer. The network is modeled as a unit‑disk graph (UDG) where nodes are placed arbitrarily in the plane and share a single communication channel. The adversary knows the protocol, the entire transmission history, and may jam any node for up to a (1‑ε) fraction of the time steps, where ε ∈ (0,1) is a constant. Crucially, nodes cannot distinguish a jammed transmission from a collision of ordinary packets, which reflects the realistic limitation of low‑cost radios.

To address this, the authors propose J‑A‑D‑E (Jamming‑Adaptive Distributed Election), a simple local MAC protocol. Each node maintains a transmission probability p that it updates autonomously based solely on its own success/failure feedback. After a successful transmission p is increased by a factor (1 + α); after a failure (which could be a jam or a collision) p is decreased by (1 − β). The parameters α and β are fixed constants chosen as functions of ε and the network density. No global synchronization, neighbor discovery, or extra control messages are required.

The theoretical analysis proceeds in two stages. First, the authors show that because the jammer can occupy at most (1‑ε) of the time slots, at least ε·T slots (over a horizon of T steps) remain “unjammed” and thus potentially usable. Second, they model the evolution of p as a Markov chain and derive concentration bounds that guarantee, with high probability, that the chain spends a constant fraction of the unjammed slots in a “high‑p” regime where successful transmissions occur. The resulting throughput per node is Θ(ε / Δ), where Δ denotes the size of a maximum independent set in the UDG (i.e., the number of nodes that can transmit simultaneously without interference). This matches the known upper bound up to constant factors, making the protocol asymptotically optimal.

When the node placement is sufficiently dense—specifically when the average inter‑node distance is at most half the communication radius—the maximum independent set size Δ becomes small, and the aggregate network throughput approaches Θ(ε·n), where n is the number of nodes. In other words, even if the jammer blocks almost the entire channel, the remaining ε fraction can be exploited almost fully.

Extensive simulations validate the analytical claims. The authors evaluate J‑A‑D‑E on two topologies (random uniform placement and regular grid) and compare it against classic ALOHA, CSMA/CA, and a recent adaptive MAC scheme. They vary ε from 0.1 to 0.5, adjust node density, and test three jammer strategies (static, periodic, and rapidly changing). Across all scenarios J‑A‑D‑E achieves a per‑node success rate between 0.25 and 0.45, which is roughly two to three times higher than the baselines. Moreover, the protocol quickly adapts when the jammer’s aggressiveness changes, keeping performance degradation minimal. In dense networks (density λ ≥ 4 · π⁻¹·r⁻², with r the radio range) the total throughput saturates near the theoretical optimum.

The paper also discusses practical deployment. J‑A‑D‑E requires only a software modification to existing radios; no extra hardware or frequency hopping is needed. The authors argue that the same probabilistic back‑off principle can be extended to multi‑channel environments. Open research directions include handling asynchronous clocks, mobile nodes, and jammers that can vary transmission power.

In summary, the work demonstrates that a purely local, probabilistic MAC protocol can achieve near‑optimal throughput in the presence of an omniscient, adaptive jammer that can dominate the channel for an arbitrary (1‑ε) fraction of time. This result is significant for security‑critical wireless sensor networks, IoT deployments, and tactical communications where robust operation under hostile interference is essential.