|
Lecture May 27 17:15 |
With the increasing complexities of systems, made possible through higher levels of integration and lower manufacturing costs, ensuring dependable operation is becoming more and more difficult. Factors affecting dependability include design bugs, manufacturing defects and process variations, environmental factors and external attacks. This lecture will focus on techniques for ensuring dependable operation of microprocessors. These include approaches for validating that the designs are free of bugs, using engineering and formal techniques, as well as methods for system design so that the processors will continue to operate in spite of failures of some components. The lecture will also address ways of tolerating external attacks on processors.
Students should have a basic understanding of logic design, computer architecture and semiconductor electronics. Any introductory university textbook on these subjects would suffice for the preliminary reading.
Students will gain a basic understanding of the fundamentals of fault-tolerant microprocessors, including design and analysis of systems based on them. Students will also learn industry practice and university research directions in these topics. Students will have the background for fault-tolerant systems in industry, or for research in the field of fault tolerance.
- Fundamentals of verification and validation
- Simulation techniques and use of assertions
- Formal equivalence checking and property checking
- Industry applications and research directions
- Redundancy techniques for fault tolerance
- Hardware and time redundancy
- System architectures for fault tolerance
- Fault-tolerant system examples
- Overview of fault-tolerant system evaluation
- Application-level techniques for fault tolerance
- Hardware and software approaches
- Security issues
- Dealing with attacks