Gunwoo Kim

About

I'm a systems and network security researcher working at the intersection of SmartNICs, RDMA, and trusted execution. I completed my M.S. at Kwangwoon University and am currently a M.S Researcher at Hanyang University, studying hardware and systems security with TEEs and SmartNICs. My recent work on RDMA resource-exhaustion attacks and defenses appeared at Network & Security 2026 (Elsevier).

Experience

• M.S. Researcher 2026.02 — Present
  SSLab, Hanyang University

  Researching hardware&systems security using Trusted Execution Environments (TEE) and SmartNICs.

• Teaching assistant 2023.03 — 2025.02
  Kwangwoon University

  Linux utilization practice, data structure, algorithms, artificial intelligence and digital logic circuits.

• High School Student Mento 2023.06 — 2024.09
  & Centor

  As a student counsellor, I offered career guidance and helped high school students develop career objectives related to the software engineering field. Career education for college entrance for high school students and Youth Hackathon Mento

• Information Systems Analyst (information protection unit) 2021.03 — 2022.09
  Republic of Korea Army, 3rd Corps, 103rd Information and Communications Group

  As an Information Systems Analysis Specialist, I managed and operated the IPS, FW, and NAC equipment for the 3rd Corps and its subordinate units. Additionally, I was responsible for managing and documenting the internal intranet network of the military. Additional tasks included packet analysis, virus analysis, and backup. (Details about the specific equipment and tasks cannot be disclosed due to a military secrecy agreement.)

Education

• M.S. in Computer Science 2023.03 — 2025.02
  Kwangwoon University, Graduate School

  Completed classes related to artificial intelligence and data science

• B.S. in Computer Science 2019.03 — 2023.02
  Kwangwoon University, Information Science Education Institute

  Completed information security-related classes and major computer engineering

Publications

• MIDAS: Mitigating Microarchitectural Resource Exhaustion in RDMA Containers via SmartNIC Under revision, 2026
  Computers & Security (SCI Journal)

  Designed novel microarchitectural resource-exhaustion attacks against RDMA NICs in containerized clouds, and proposed an RNIC-resident defense framework combining EWMA-CUSUM anomaly detection on ARM cores with an INT8-quantized LSTM-GRU classifier offloaded to the Data Path Accelerator.

• Noisy Neighbor: Exploiting RDMA for Resource Exhaustion Attacks in Containerized Clouds Sep 2025
  SecAssure Workshop at ESORICS 2025

  Analyzed RDMA resource-exhaustion attacks in containerized clouds on NVIDIA BlueField-3, identifying state and pipeline saturation vulnerabilities that degrade co-tenant performance (up to −93.9% bandwidth, +1,117× latency). Proposed HT-Verbs, a threshold-based framework that classifies RDMA workloads by activity level to mitigate abuse.

• M.S Thesis: Design and Implementation of Tenant Resource Orchestration Based on Multi-Policy Reinforcement Learning Feb 2025
  Master Thesis at Kwangwoon University
• Analysis of RDMA SmartNIC Resource Exhaustion in Container Environments Feb 2025
  Proceedings of Symposium of the Korean Institute of communications and Information Sciences
• Research on Hot-Threshold Based Dynamic Resource Management in the Cloud Sep 2024
  International Journal of Advanced Culture Technology
• A Study on Cold-Start and Resource Improvement Using Time Warming Allocation Engine in Serverless Computing Sep 2024
  International Journal of Advanced Culture Technology
• An Engine for DRA in Container Orchestration Using Machine Learning Dec 2023
  International Journal of Advanced Culture Technology

Projects

• Comparative Analysis of Data-Driven Scheduling Algorithms in Serverless Computing 2024.03 — 2024.06
  Kwangwoon University, Graduate School

  Compared scheduling algorithms (FIFO, SEPT, FCFS variants, RR, SERPT) under Azure-trace constraints to reduce response times in serverless computing.

• University Lab Management Program 2024.01 — 2024.02
  Kwangwoon University, Graduate School

  Built a program using the Windows API and C++/CLI that blocks unnecessary functions to create a focused study environment for students.

• Hacking using Arduino Module 2019 · 2023
  Academic Festival & Nowon Youth Joint Festival

  Demonstrated network and hardware exploits (diskpart abuse, data extraction, webcam control) and simulated a router hijack using an Arduino network module and Raspberry Pi to raise cybersecurity awareness.

• 3rd Corps Network Control Assistance Program 2021.01 — 2022.08
  Republic of Korea Army, 3rd Corps

  Developed a web application that documented threat situations and anomalies from equipment data, with automated analysis via shell scripts and C++.

• Security Incident Prevention Game 2021.05 — 2021.08
  Republic of Korea Army, 3rd Corps

  Created a 2D pixel game (HTML, CSS, JavaScript, design patterns; no external APIs) teaching network-security incident prevention — 2nd place in an ROK Army competition and featured on the military intranet.

Skills

• Confidential Computing
• SmartNIC / BlueField
• RDMA / RoCEv2
• Network Security
• Machine Learning for Systems
• C / C++
• Python

Certificates

• AWS Cloud Practitioner
• Engineer Information Processing