Fault-Tolerance and Performance in Real-Time Systems

当前的实时系统通常是在这样的假设下设计的：在系统设计时，工作负载和计算环境的特征都是已知的，并且都是明确的。然而，未来的实时系统将需要在更加不确定和随机的环境中运行。因此，这些未来系统的设计者在开发资源调度算法和正式建立它们的性能、定时和容错特性方面面临着更加困难的任务。本课题关注随机环境下实时系统调度算法的设计和性能评估。具体的重点是开发严格的和正式的方法来分析和设计这样的系统在广泛的设置，包括单处理器，多处理器和容错环境。

Current real-time systems have typically been designed under the assumption that characteristics of the workload and the computing environment are both well-known and well-specified at the time of system design. Future real-time systems, however, will be required to operate in a more uncertain and stochastic environment. As a result, the designers of these future systems face a much more difficult task in developing resource scheduling algorithms and formally establishing their performance, timing, and fault-tolerance properties. This project is concerned with the design and performance evaluation of scheduling algorithms for real-time systems operating in a stochastic environment. The specific focus is on developing rigorous and formal approaches towards analyzing and designing such systems in a wide range of settings including uniprocessor, multiprocessor and fault-tolerant environments.