Rock Hyun Choi
MOEM Lab in Daegu
University

 What is CPS?
The integration of physical system and processes with networked computing has led to the emergence of a new generation of engineered system

 Range
 Miniscule(pace makers) ~ large scale(Aircraft)
 Good Ex:
 Automobiles: the depth of this change is wellreflected
 Why important?
 다른 방법으로 실현 할 수 없었던 것을 임베디드
시스템을 이용하면 physical system에 어떤 능력
을 부여해 줄 수 있다.
 Ex) fuel efficiency, pollution mandates

 How to make it better?
 Merging computing and networking with physical system
 So, it creates new capabilities and improve/maintain product quality.
 Methodology
 U.S. – NSF
 European Union – ARTEMIS & EPoSS
 JAPAN – Embedded technology conference
 South Korea - KIPA

 Opportunities Created by CPS


Opportunities Created by CPS

 Shortcomings
 Current industrial experience
→ inefficient, unsafe, and unpredictable
 Composition
 Compositionality: 시스템 레벨 영역들을 로컬영역
구성요소들에 의해서 계산(조작) 될 수 있다.
 Composability: 컴포넌트(부품, 성분, 구성요소)
영역은 다른 컴포넌트에 상호작용의 결과로 변화
되지 않는다.

 Composition
 CPSs are inherently heterogeneous
 Why? Their components and essential design requirements.
 Multi-objective design problems:
 Separation of concerns( = design viewpoints)
Engine engineer
Network engineer orthogonal

 Design automation
 Heterogeneity → increase design complexity
• → bad for design automation, serious problem
 Today standard is to use different processes and tools
• 중복된 노력으로 시간과 자원 낭비가 심함

 System integration into integrate complex components
Complex
System
• Cannot reliable, predictable
• Large scale system development에 일반적 문제
• Why? Science 와 engineering의 분리로 발생한 관리
문제

 Certification
 현재 complex system(ex. 비행기, 의료장비)
인증비용이 너무 높고, 이전 디자인 양상으로
는 모두 역어서 테스트 하기가 너무나 힘듬
First building
And then check
검
사
기

 Security and privacy
 It’s new threats!
 Can cause huge accidents → can’t easily recover
 Why? Physical system can now be attacked through cyberspace, and vice versa.
 Education
 Stovepiping of engineering versus computer science
• Fail to train CPS workers

 The challenges of CPS research
 2.1 Real-time system abstractions
• Current: the principal task of a computer is data transformation
• CPS: CPSs are real-time system. This requires reexamination of existing hardware and software architecture.
– Real-time event triggers
– Consistent views of distributes states in real-time within the sphere of influence. Great in mobile devices
– Topology control and dynamic real-time groups in the form of packaged service classes of bounded delay, jitter and loss under precisely specified condition.
– Interface to access to the same type of controls regardless of the underlying network technology

 임베디드 시스템은 다른 일반 목적 컴퓨팅 시
스템 보다 높은 안정성을 추구한다
 Not entirely predictable
 Why? It misses timing deadline
1.
Worst Case Execution Time(WCET)
• 간단한 문제해결을 위해서도 많은 노력이 필요
2.
RTOS
• Timing behavior → coarse and complexity → more uncontrollable
• Bench testing 필요 : 그러나 포작하기 어려운 미묘한 timing bugs 가 큰 문제를 일으킬 수 있음

 Solution
 Timing이 고려된 언어들
• Ex)spilit-C, Cik, SHIM
• Guava language – 멀티플 쓰레드에서 비동기화된
객체를 자바 오브젝트에서 제한시킴
• Giotto
 그 밖에 다양한 연구들이 실시간 시스템을 위
해 연구된 논문이 나와있음