Wireless Networks under a Backoff Attack: A Game Theoretical Perspective.

Area of Science:

• Computer Science
• Network Security
• Game Theory

Background:

• Wireless sensor networks (WSNs) commonly use Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) for medium access control.
• A backoff attack, where some nodes deviate from the protocol, can compromise network performance and fairness.
• Understanding the impact of such attacks is crucial for robust WSN design.

Purpose of the Study:

• To analyze the effect of backoff attacks on WSN throughput and fairness.
• To model the attacker-defender interaction using game theory.
• To develop and validate an algorithm for achieving network equilibrium under attack.

Main Methods:

• Utilized Bianchi's network model to analyze throughput under attack conditions.
• Applied game theory concepts, treating each sensor as a player in a strategic interaction.
• Developed a Regret Matching algorithm to find the game's equilibrium for multiple players.
• Validated the theoretical findings through network simulations.

Main Results:

• Demonstrated that backoff attacks lead to unfair distribution of network throughput.
• Derived analytical solutions for the game model representing the network conflict.
• The proposed Regret Matching algorithm effectively learns the equilibrium strategy.
• Simulation results confirmed the accuracy of the theoretical predictions.

Conclusions:

• Backoff attacks significantly impair fairness in CSMA/CA based WSNs.
• Game theory provides a powerful framework for analyzing such security vulnerabilities.
• Regret Matching offers a viable algorithmic solution for mitigating attack impacts and restoring network balance.