Cache performance in NDN networks with ADMM algorithm to deal with pollution attacks and traffic load changes
atefeh vaez shahrestani, Mohammad Reza Khayyambashi, Faramarz Safi · Journal of Artificial Intelligence in Electrical Engineering · 2025
As data-driven networks like Named Data Networks (NDN) continue to grow, securing cache systems has become an increasingly critical issue, especially in defending against cache pollution attacks (CPA). This paper introduces a novel algorithm based on the Alternating Direction Method of Multipliers (ADMM) method for adaptively tuning the β parameter in the PFP-β-DA cache policy. The algorithm works by decomposing the optimization problem into manageable subproblems and updating the parameters in a distributed fashion. This approach enhances the cache CHR and reduces cache pollution (CPR). The algorithm was tested in three network topologies—simple, hierarchical, and advanced—and evaluated under stable conditions, attack scenarios, and dynamic traffic. Simulation results demonstrate that ADMM outperforms traditional methods like LRU, LFU-DA, and PFP-β-DA, providing better stability, adaptability, and precision. Particularly, in high-attack scenarios, the proposed algorithm reduces cache pollution by as much as 5% and increases the cache hit rate by up to 4% over the baseline methods. These results confirm the effectiveness of ADMM as a scalable and intelligent solution for improving cache security and efficiency in NDN networks.