On the Routing Protocol Influence on the Resilience of Wireless Sensor Networks to Jamming Attacks

Juan Arturo Nolazco Flores, Carolina Del-Valle-Soto, J. Carlos Mex-Perera, Raúl Monroy, Juan Arturo Nolazco-FloresSensors 2015 · DOI

TL;DR

This paper investigates how different routing protocols (AODV, DSR, and MPH) affect the resilience of Wireless Sensor Networks (WSNs) against jamming attacks. The study found that the Multi-Parent Hierarchical (MPH) protocol significantly outperforms AODV and DSR in terms of packet retransmissions, CSMA retries, and energy consumption when subjected to jamming, due to its ability to isolate attacked segments and its proactive-reactive nature.

What Problem Does This Paper Address?

Wireless Sensor Networks (WSNs) are increasingly deployed for critical applications, but they are vulnerable to various threats, including jamming attacks. Jamming aims to disrupt communication by intentionally interfering with radio signals. While WSNs are designed to be robust, the effectiveness of their resilience mechanisms, particularly how routing protocols handle such attacks, is not fully understood. This paper addresses the need to evaluate and compare the resilience of common WSN routing protocols against different types of jamming attacks, focusing on how these protocols adapt and maintain network performance under duress. The authors aim to identify which routing protocol offers superior resilience and better overall network performance when faced with jamming.

Methods

The authors developed a simulator to faithfully replicate the workings of three routing protocols: Ad hoc On-Demand Distance Vector (AODV), Dynamic Source Routing (DSR), and the proposed Multi-Parent Hierarchical (MPH) protocol. They simulated two types of jamming attacks: random and reactive, considering various jammer positions. The performance of the protocols was evaluated using several metrics:

  • Number of packet retransmissions: Indicates link quality and network connectivity.
  • CSMA retries: Measures the number of times a node attempts to transmit a packet when the channel is busy, reflecting wireless medium congestion.
  • Tries metric: Represents the number of attempts by the coordinator to communicate with each node, indicating route availability and network topology knowledge.
  • Achievable nodes by the coordinator: The number of nodes that can respond to the coordinator's queries, reflecting the network's ability to maintain connectivity.
  • Energy consumption: Quantifies the resource usage of each node.

The study employed box and whisker plots for statistical analysis and 2D spatial distribution figures to visualize the impact of jamming across the network topology.

Datasets

Not discussed in the paper. The study relies on a custom-built simulator and simulated network scenarios rather than real-world datasets.

Key Insights and Ideas

  • MPH's localized attack handling: The MPH protocol's ability to "encapsulate" or isolate the network segment under attack is a key factor in its superior resilience. This prevents the attack from spreading and affecting a larger portion of the network.
  • Proactive-reactive nature of MPH: MPH's combination of periodic neighbor table updates (proactive) and reactive mechanisms allows it to adapt quickly to link failures caused by jamming, while also maintaining efficient routes.
  • Jammer position matters: The impact of jamming is significantly influenced by the jammer's location. Attacks closer to the coordinator node tend to have a more disruptive effect on overall network performance.
  • AODV/DSR's vulnerability: Reactive protocols like AODV and DSR, which rely on route discovery when needed, experience increased overhead and packet loss when the topology is constantly changing due to jamming, making them more susceptible to attack propagation.
  • Self-configuration as a resilience pillar: The study highlights that effective self-configuration mechanisms in routing protocols are crucial for WSN resilience, enabling rapid link recovery and topology adaptation.

Key Results

  • Retransmissions: MPH consistently showed significantly fewer retransmissions (e.g., 13-30% less than AODV/DSR) across various jamming scenarios, indicating better link stability.
  • CSMA Retries: MPH exhibited lower CSMA retries (e.g., 18.4% less than AODV/DSR for random jamming with p=50, and up to 61% less for reactive jamming when the jammer is far) compared to AODV and DSR, suggesting less channel congestion and fewer collisions.
  • Tries: MPH required fewer tries for the coordinator to reach nodes, indicating better route availability and knowledge of the network topology (e.g., 15.8% better connectivity than DSR and 19.4% better than AODV in no-jamming scenarios).
  • Achievable Nodes: MPH maintained a higher number of reachable nodes by the coordinator, demonstrating its superior self-configuration and ability to recover broken links.
  • Energy Consumption: MPH demonstrated significant energy savings (e.g., 42% compared to AODV and 37% to DSR in no-jamming scenarios, and up to 46% savings for nodes near the coordinator during reactive jamming).

Limitations and Future Work

  • Jamming Detector Construction: The study notes that while MPH's consistent performance under jamming is desirable, it might make the construction of a jamming detector more challenging compared to AODV and DSR, where metric deviations are larger. Further metrics might be needed for specific protocols.
  • Improving Reaction to Jamming: The authors suggest improving the reaction mechanism for jamming attacks. Specifically, they propose using a larger history of neighbor activity to refine the persistence counter for neighbor steadiness, which could reduce false positives in node removal but might increase reconfiguration delay.
  • Multi-feature Detector/Protocol: Future work could involve developing a multi-feature detector or a routing protocol with enhanced capabilities to deal with jamming attacks, potentially by integrating signal strength and other MAC/network layer parameters.

Summary updated

BibTeX
@article{flores2015routing,
  author = {Juan Arturo Nolazco Flores and Carolina Del-Valle-Soto and J. Carlos Mex-Perera and Raúl Monroy and Juan Arturo Nolazco-Flores},
  title = {{On the Routing Protocol Influence on the Resilience of Wireless Sensor Networks to Jamming Attacks}},
  journal = {Sensors},
  year = {2015},
  doi = {10.3390/s150407619},
}
Figure