CAREER: Towards a rigorous foundation for scheduling in modern systems

点击翻译按钮获取中文摘要

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).Scheduling policies are at the heart of all computer systems and are a key determinant of system performance. The goal of this project is to provide a rigorous foundation for modern scheduling design issues. The project includes two main directions that correspond to two important modern design paradigms. The first direction is understanding the impact of prioritization. Prioritization is now a fundamental part of system design, and it is applied for many reasons, e.g., to provide QoS guarantees, to provide differentiated service, or simply to provide improved performance. The second direction is understanding scheduling in distributed/parallel architectures.Distributed/parallel designs are now the norm rather than the exception, and they present a wide variety of important scheduling and resource allocation issues.Across these two directions there are two themes that play a prominent role in the research. The first theme is the importance of power management in modern designs. As energy costs soar, power management is increasingly being treated as a first-class design metric, and must be considered when designing scheduling policies. The second theme is the benefit of applying economic tools, such as game theoretic techniques, to approach scheduling questions. Economic tools are increasingly being exploited with great success by computer scientists, and the domain of scheduling is no exception.Because scheduling is important to a wide variety of scientific, engineering, computing, and business applications, this project will have a broad impact beyond computer science. However, because of the interdisciplinary nature of scheduling, there are a wide variety of approaches across very distinct communities. One goal of this project is to provide courses and educational tools that help to bring these communities together.