상세 보기
온디바이스 AI 환경에서 보안 통신 프로토콜 및 암호화 구성에 따른 딥러닝 추론 성능 영향 분석
- 채주혁;
- 권오빈;
- 김진홍;
- 안형태
초록
As on-device AI and IoT services are increasingly deployed in industrial environments, edge devices are required to process video and sensor data in real time while ensuring secure data transmission. However, limited CPU and memory resources can lead to resource contention between deep-learning inference and secure communication processing, resulting in system performance degradation. This paper provides a quantitative analysis of how various combinations of secure communication protocols and TLS 1.3 cipher suites impact inference speed(FPS), network latency(RTT), and CPU/memory utilization in on-device AI environments. Based on these empirical results, we propose practical guidelines for selecting protocol and cryptographic configurations that balance real-time performance with security requirements. These findings offer foundational insights for designing future lightweight secure communication architectures for on-device AI.
키워드
- 제목
- 온디바이스 AI 환경에서 보안 통신 프로토콜 및 암호화 구성에 따른 딥러닝 추론 성능 영향 분석
- 제목 (타언어)
- Performance Impact of Secure Communication Protocols and Cryptographic Configurations on On-Device AI Inference
- 저자
- 채주혁; 권오빈; 김진홍; 안형태
- 발행일
- 2026-04
- 유형
- Y
- 저널명
- 한국정보기술학회논문지
- 권
- 24
- 호
- 4
- 페이지
- 73 ~ 84
- 언어
- KOR
- 출판사
- 한국정보기술학회
- 발행국가
- 대한민국
- 분량
- 12 페이지
- ISSN
- E 2093-7571
P 1598-8619