ADAPTIVE ATTACK-RESISTANT ALGORITHM AND SECURITY ORCHESTRATION FRAMEWORK FOR NAMED-DATA NETWORKING

Abstract

Named-Data Networking (NDN) is a concept of delivering a paradigm shift between host-centric and content-centric communication, which facilitates an efficient in-network-based caching and name-based data-retrieval. Nevertheless, its architectural elements, Content Store (CS), Pending Interest Table (PIT) and Forwarding Information Base (FIB) provoke peculiar security weaknesses, especially Interest Flooding Attacks (IFA), cache poisoning, and content forgery. Such attacks affect the availability of the network, utilize PIT resources, and affect data integrity. The present paper will present an adaptive attack-resistant algorithm together with a security orchestration framework that will add resilience to the NDN setting. The proposed method integrates machine learning-based real-time attack classification under supervision with dynamic mitigation using reinforcement learning in an attempt to implement better rate limiting, PIT threshold control, and cache validation policies in a dynamically adjusted fashion. A cross-layer orchestration layer, which is a coordination layer between security decisions in CS, PIT, and FIB modules, is a distributed controller to provide a unified policy enforcement and quick reaction. To model the dynamics of PIT occupancy, attack detection, probability of false positive and latency throughput trade-offs are represented using mathematical models in adversarial conditions. The outcomes of the simulation reveal that there are tremendous gains in the detection accuracy, and the false positive rates are lowered, the usage of PIT has been stabilized, and the latency increase is minimized in relation to traditional, non-adaptive defense mechanisms.