Repeated Game Analysis of a CSMA/CA Network under a Backoff Attack

Area of Science:

• Computer Science
• Network Engineering
• Game Theory

Background:

• Wireless networks often use Carrier Sense Medium Access with Collision Avoidance (CSMA/CA) for channel access.
• Deviations from the standard CSMA/CA protocol by some stations can negatively impact network performance and fairness.
• Previous work modeled this scenario using static game theory.

Purpose of the Study:

• To analyze the impact of protocol deviations in CSMA/CA wireless networks.
• To investigate the use of repeated game theory to improve outcomes for all network stations.
• To develop and validate analytical and distributed algorithms for learning equilibrium strategies.

Main Methods:

• Utilizing Bianchi's model to analyze network throughput and fairness.
• Applying repeated game theory, including Folk theorem concepts.
• Deriving analytical solutions for two-player scenarios using subgame perfect and correlated equilibria.
• Proposing a distributed algorithm for learning equilibria with multiple players.
• Conducting numerical simulations for validation and comparison.

Main Results:

• Protocol deviations in CSMA/CA networks significantly reduce network fairness.
• Deviating stations gain a disproportionately larger share of network resources.
• Repeated game theory offers potential for improved outcomes compared to static game models.
• Analytical and distributed algorithms provide methods for identifying and achieving equilibria.

Conclusions:

• Deviations in CSMA/CA protocols severely compromise network fairness.
• Repeated game theory and equilibrium concepts offer viable solutions for mitigating unfairness.
• The proposed distributed algorithm is effective for learning equilibria in multi-player scenarios.
• Numerical simulations confirm the effectiveness of the proposed analytical and algorithmic approaches.