Sergio Elizalde
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Sergio Elizalde has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Elie Kfoury
University of South Carolina
4 shared publications - Jorge Crichigno
University of South Carolina
4 shared publications - Ali AlSabeh
University of South Carolina Aiken, Aiken, USA
3 shared publications - Ali Mazloum
University of South Carolina, Columbia, South Carolina, USA
3 shared publications - Jose Gomez
Fort Lewis College
3 shared publications - Samia Choueiri
University of South Carolina, Columbia, USA
3 shared publications
Latest Publications
- Open Access
- 19 September 2026
- Published by Elsevier in SSRN Electronic Journal
SmartNIC-Based Network Traffic Monitoring: Taxonomy, Offloading Models, and Open Challenges
The scale, speed, and complexity of modern network infrastructures are placing increasing pressure on traditional traffic monitoring architectures based on host-side processing, periodic polling, and centralized collection. At the same time, high-speed environments such as cloud platforms, data centers, enterprise networks, and edge systems demand fine-grained visibility for performance management, troubleshooting, traffic analytics, and security. Smart Network Interface Cards (SmartNICs) have emerged as promising platforms for moving telemetry and traffic analysis closer to the data path, enabling low-latency and scalable monitoring while reducing host-CPU overhead. Research in this area, however, remains dispersed across telemetry systems, programmable packet processing, traffic analytics, security mechanisms, and cloud or edge acceleration platforms, making it difficult to understand how monitoring functions are partitioned across heterogeneous devices and processing tiers. This paper presents a monitoring-oriented review of SmartNIC-based traffic monitoring systems. It introduces a function-driven taxonomy covering performance monitoring, flow tracking, traffic analytics, and threat detection and mitigation, and analyzes these systems through four cross-cutting offloading models that describe the role actually played by the SmartNIC within the monitoring pipeline. The resulting corpus-level analysis relates monitoring function, execution substrate, offloading strategy, and evaluation evidence, revealing recurring architectural patterns and substantial gaps between prototype implementation and operational deployment. Finally, the survey derives open challenges and research directions concerning state management, cross-tier coordination, portability, encrypted-traffic monitoring, trustworthiness, energy-aware placement, and evaluation realism.
- 19 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSmartNIC-Based Network Traffic Monitoring: Taxonomy, Offloading Models, and Open Challenges
The scale, speed, and complexity of modern network infrastructures are placing increasing pressure on traditional traffic monitoring architectures based on host-side processing, periodic polling, and centralized collection. At the same time, high-speed environments such as cloud platforms, data centers, enterprise networks, and edge systems demand fine-grained visibility for performance management, troubleshooting, traffic analytics, and security. Smart Network Interface Cards (SmartNICs) have emerged as promising platforms for moving telemetry and traffic analysis closer to the data path, enabling low-latency and scalable monitoring while reducing host-CPU overhead. Research in this area, however, remains dispersed across telemetry systems, programmable packet processing, traffic analytics, security mechanisms, and cloud or edge acceleration platforms, making it difficult to understand how monitoring functions are partitioned across heterogeneous devices and processing tiers. This paper presents a monitoring-oriented review of SmartNIC-based traffic monitoring systems. It introduces a function-driven taxonomy covering performance monitoring, flow tracking, traffic analytics, and threat detection and mitigation, and analyzes these systems through four cross-cutting offloading models that describe the role actually played by the SmartNIC within the monitoring pipeline. The resulting corpus-level analysis relates monitoring function, execution substrate, offloading strategy, and evaluation evidence, revealing recurring architectural patterns and substantial gaps between prototype implementation and operational deployment. Finally, the survey derives open challenges and research directions concerning state management, cross-tier coordination, portability, encrypted-traffic monitoring, trustworthiness, energy-aware placement, and evaluation realism.
- Open Access
- 15 June 2026
- Published by Elsevier in SSRN Electronic Journal
A Comprehensive Survey of Acceleration Techniques for Intrusion Detection and Prevention Systems: Taxonomy, Architectures, Challenges, and Future Trends
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.
- 15 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Comprehensive Survey of Acceleration Techniques for Intrusion Detection and Prevention Systems: Taxonomy, Architectures, Challenges, and Future Trends
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.
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 860-865
Optimizing Network Resilience Using Domain-Specific Hardware Accelerator for Dynamic Programming
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 860-865
Open AccessOptimizing Network Resilience Using Domain-Specific Hardware Accelerator for Dynamic Programming
- Open Access
- 2 July 2025
- Published by Elsevier in Journal of Network and Computer Applications
- Vol. 242
A survey on security applications with SmartNICs: Taxonomy, implementations, challenges, and future trends
- 2 July 2025
- Published by Elsevier in Journal of Network and Computer Applications
- Vol. 242
Open AccessA survey on security applications with SmartNICs: Taxonomy, implementations, challenges, and future trends