MPH-M, AODV-M and DSR-M Performance Evaluation under Jamming Attacks

Carolina Del-Valle-Soto, Carlos Mex-Perera, Raúl Monroy, Juan A. Nolazco‐FloresSensors 2017 · DOI

TL;DR

This paper introduces a novel two-phase jamming attack mitigation scheme for Wireless Sensor Networks (WSNs) integrated into existing routing protocols (MPH, AODV, DSR). The proposed MPH-M protocol demonstrates significantly improved performance in terms of energy efficiency, throughput, and resilience compared to modified AODV-M and DSR-M, especially under jamming conditions.

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 can saturate receivers with spurious signals, leading to dropped packets, increased retransmissions, and ultimately, a denial of service. Existing routing protocols often struggle to cope with these attacks, leading to network degradation and wasted energy. This paper addresses the challenge of maintaining WSN functionality and performance in the face of jamming by developing and evaluating a robust mitigation strategy.

Methods

The core of this work is a two-phase mitigation mechanism:

  1. Detection: Each node locally monitors key performance metrics – specifically, energy consumption, packet retransmissions, and CSMA/CA retries. An anomaly detection model, trained on normal network behavior, identifies deviations that indicate a jamming attack. A majority voting rule using these three metrics helps reduce false positives.
  2. Reconfiguration: Upon detecting a jamming attack, a node sends a control packet to its neighbors. This triggers a route reconfiguration process where the jammed node is isolated and removed from routing tables. The routing protocols then adapt their paths to avoid the affected area.

The authors modified three well-known routing protocols:

  • MPH (Multi-Parent Hierarchical): Modified to MPH-M.
  • AODV (Ad hoc On Demand Distance Vector): Modified to AODV-M.
  • DSR (Dynamic Source Routing): Modified to DSR-M.

These modified protocols were then evaluated under various jamming scenarios.

Datasets

Not discussed in the paper. The authors used a custom event-driven simulator based on C++ for their experiments.

Key Insights and Ideas

  • Proactive Protocols Shine: The proactive nature of the MPH protocol, when modified to MPH-M, appears to be inherently more resilient to jamming attacks than reactive protocols like AODV and DSR. This is likely due to its hierarchical structure and more stable link establishment.
  • Combined Metrics are Key: Relying on a single metric for jamming detection can lead to false positives. The paper demonstrates that combining energy consumption, retransmissions, and CSMA/CA retries provides a more robust detection mechanism.
  • Mitigation Benefits MPH Disproportionately: While all modified protocols show improvement with the mitigation scheme, MPH-M benefits the most, suggesting that its underlying architecture is better suited to adapt to jamming-induced disruptions.
  • Jammer Location Matters: The impact of a jamming attack is significantly influenced by the jammer's position relative to the collector node. Attacks closer to the collector have a more severe impact on throughput.

Key Results

  • Energy Efficiency: MPH-M shows substantial energy savings, being 138.13% better than AODV-M and 126.07% better than DSR-M in terms of node energy. The mitigation scheme further reduces energy consumption by 34.61% for MPH-M.
  • Throughput: MPH-M achieves higher packet reception efficiency, outperforming AODV and DSR by 16.4% even without mitigation. With mitigation, MPH-M shows an 8.4% efficiency improvement, while AODV-M and DSR-M show negligible improvements.
  • Detection Accuracy: MPH-M demonstrates superior jamming detection, achieving up to 100% efficiency in identifying affected nodes, with significantly fewer false alarms (1.75%) compared to AODV-M (8.25%) and DSR-M (9.0%).
  • Resilience: MPH-M exhibits a 15% lower packet retransmission change rate and a faster recovery time (5s faster than AODV-M, 2s faster than DSR-M) after a jamming attack.

Limitations and Future Work

  • Fixed Topology: The experiments were conducted on a static grid topology. Future work could explore the performance in dynamic or mobile network scenarios.
  • Specific Jamming Model: The study focused on a reactive jamming model. Evaluating performance against other jamming types (e.g., constant, deceptive) would be beneficial.
  • Protocol Comparison: The authors mention other protocols like RPL and modified OSLR, suggesting future work could involve comparing MPH-M with these as well.
  • Jammer Localization: While the paper focuses on detection and mitigation, jammer localization is a related area that could be explored further.

Summary updated

Abstract

In this work, we present the design of a mitigation scheme for jamming attacks integrated to the routing protocols MPH, AODV, and DSR. The resulting protocols are named MPH-M (Multi-Parent Hierarchical - Modified), AODV-M (Ad hoc On Demand Distance Vector - Modified), and DSR-M (Dynamic Source Routing - Modified). For the mitigation algorithm, if the detection algorithm running locally in each node produces a positive result then the node is isolated; second, the routing protocol adapts their paths avoiding the isolated nodes. We evaluated how jamming attacks affect different metrics for all these modified protocols. The metrics we employ to detect jamming attack are number of packet retransmissions, number of CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance) retries while waiting for an idle channel and the energy wasted by the node. The metrics to evaluate the performance of the modified routing protocols are the throughput and resilience of the system and the energy used by the nodes. We evaluated all the modified protocols when the attacker position was set near, middle and far of the collector node. The results of our evaluation show that performance for MPH-M is much better than AODV-M and DSR-M. For example, the node energy for MPH-M is 138.13% better than AODV-M and 126.07% better than DSR-M. Moreover, we also find that MPH-M benefits much more of the mitigation scheme than AODV-M and DSR-M. For example, the node energy consumption is 34.61% lower for MPH-M and only 3.92% and 3.42% for AODV-M and DSR-M, respectively. On throughput, the MPH protocol presents a packet reception efficiency at the collector node of 16.4% on to AODV and DSR when there is no mitigation mechanism. Moreover, MPH-M has an efficiency greater than 7.7% with respect to AODV-M and DSR-M when there is a mitigation scheme. In addition, we have that with the mitigation mechanism AODV-M and DSR-M do not present noticeable modification. However, MPH-M improves its efficiency by 8.4%. We also measure the resilience of these algorithms from the average packet re-transmissions perspective, and we find that MPH-M has around a 15% lower change rate than AODV-M and DSR-M. The MPH-M recovery time is 5 s faster than AODV-M and 2 s faster than DSR-M.
BibTeX
@inproceedings{delvallesoto2017mph,
  author = {Carolina Del-Valle-Soto and Carlos Mex-Perera and Raúl Monroy and Juan A. Nolazco‐Flores},
  title = {{MPH-M, AODV-M and DSR-M Performance Evaluation under Jamming Attacks}},
  booktitle = {Sensors},
  year = {2017},
  doi = {10.3390/s17071573},
}
Figure