NOS-Gate: Queue-Aware Streaming IDS for Consumer Gateways under Timing-Controlled Evasion
Researchers have developed NOS-Gate, a new intrusion detection system designed for consumer gateways that operates on streaming metadata from encrypted traffic. This system uses a lightweight spiking neural network model to score traffic windows and can temporarily reduce a flow's weight under weighted fair queueing to mitigate threats. Evaluations show NOS-Gate achieves high incident recall with a low false-positive rate, outperforming baseline methods while minimizing queueing delays and computational costs. AI
IMPACT Introduces a novel approach to network security using AI for real-time threat detection in constrained environments.