A Comprehensive Survey of Acceleration Techniques for Intrusion Detection and Prevention Systems: Taxonomy, Architectures, Challenges, and Future Trends
Preprint
- 15 June 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
The rapid growth of network traffic volume, link speed, and encryption has made real-time intrusion detection and prevention increasingly difficult in modern networks. Intrusion Detection Systems (IDS) and Intrusion Prevention Systems (IPS) are widely deployed to inspect traffic, identify and block malicious activity before reaching protected resources. However, conventional IDS/IPS architectures often rely on general-purpose CPUs and the traditional network stack for packet acquisition, packet inspection, flow tracking, cryptographic processing, and signature matching. As port speeds continue to outpace CPU performance, these systems face packet loss, increased latency, and missed detections, especially for encrypted traffic, stateful inspection, and inline prevention. To address these limitations, researchers and industry have proposed diverse software and hardware acceleration techniques. While prior surveys examine IDS accuracy, machine learning, and packet-processing acceleration in isolation, the literature lacks a view connecting acceleration technologies to the IDS/IPS functions they improve and the deployment trade-offs.This paper surveys acceleration techniques for IDS/IPS systems by presenting the fundamentals of detection methods, comparing different architectures, and discussing performance bottlenecks. It compares related surveys to identify the research gap, then introduces a taxonomy that maps acceleration techniques to core IDS/IPS functions. For each category, representative systems are analyzed in terms of performance, programmability, scalability, detection capabilities, and deployment implications. Finally, the survey discusses open challenges and future research directions to guide the design of scalable, high-performance network defenses.Keywords
This publication has 39 references indexed in Scilit:
- In-network computing-based malicious traffic filtering for multi-tenant cloud environmentsComputers & Security, 2026
- Line rate botnet detection with SmartNIC-embedded feature extractionComputer Networks, 2024
- A Survey on Intrusion Detection and Prevention SystemsSN Computer Science, 2023
- ThunderSecure: deploying real-time intrusion detection for 100G research networks by leveraging stream-based features and one-class classification networkInternational Journal of Information Security, 2022
- A comprehensive deep learning benchmark for IoT IDSComputers & Security, 2022
- A survey on security applications of P4 programmable switches and a STRIDE-based vulnerability assessmentComputer Networks, 2022
- The Actual Cost of Programmable SmartNICs: Diving into the Existing Limits2021
- Detection of zero-day attacks: An unsupervised port-based approachComputer Networks, 2020
- FPGA-Based Acceleration of Pattern Matching in YARA2016
- GENI: A federated testbed for innovative network experimentsComputer Networks, 2014