Student Research Paper
Dr. Jingwen Wang; Dr. Peilong Li
Detecting malicious network traffic in real time has become a crucial requirement at smart communities for elderly care and medical facilities with the prevalence of Internet-of-things (IoT) devices. Existing machine learning based solutions for network traffic malware detection often fail to scale with the exponential increase of IoT devices at the facility and to detect malicious traffic with desirable low latency. In this paper we seek to fill the gap by designing a scalable end-to-end network traffic analyzing system that permits real-time malware detection. By leveraging distributed systems such as Apache Kafka and Apache Spark, the system has demonstrated scalable performance as the number of IoT devices grow. Using Intel’s oneAPI software stack for both machine learning and deep learning models, the model inference speed is boosted by three-fold.
Weitkamp, Ethan; Satani, Yusuke; Li, Peilong; and Wang, Jingwen, "NetSec: Real-time and Scalable Malware Traffic Detection within IoT Networks" (2022). Summer Scholarship, Creative Arts and Research Projects (SCARP). 43.
Scholarship, Creative Arts, and Research Project (SCARP)