Workshop on Feedback Control of Computing Systems

计算机系统反馈控制研讨会由美国国家科学基金会于2005年5月资助，汇集了计算机系统控制领域的重要学术和工业研究人员。除了提出新的研究问题和确定当前理论和实践中的差距之外，研讨会的目标是评估该领域最近的关键贡献和研究方向的可行性、重要性、潜力和预期影响。本次研讨会的动机有两点。首先是近十年来在计算机系统控制方面取得的进展。在不同的研究领域加快了研究工作，包括网络、混合控制、实时和嵌入式计算、服务质量、Web架构、数据库、分布式系统等。第二个需要考虑的问题是，软件从业者可以在商业产品的开发中有效地使用控制理论，例如IBM在DB2中的自主特性方面的经验。研讨会寻求该领域的统一观点，以确定孤立研究问题之间的共性，提供概括的机会，为这类新的控制问题（受控过程是一个（典型的分布式）软件实体）发展广泛的理论基础，并确定可能的应用领域以产生实际影响。研讨会提供了一个跨学科交流的机会，以确定关键的研究问题，这些问题是基本的，交叉的，并且在软件系统控制领域中最广泛适用的。结果汇编将通过控制和计算机科学杂志和通讯上发表的报告向公众提供和传播。

Workshop on Feedback Control of Computing Systems (Tilbury)A Workshop on Feedback Control of Computing Systems is supported by NSF in May, 2005, to bring together key academic and industrial researchers in the field of control of computing systems. In addition to suggesting new research problems and identifying gaps in current theory and practice, the workshop goal is to assess viability, importance, potential, and expected impact of key recent contributions and research directions in the field.This workshop is motivated by two considerations. The first is progress made in control of computing systems in the current decade. Accelerated research efforts were conducted in different research communities including Networking, Hybrid Control, Real-time and Embedded Computing, Quality-of-Service, Web Architecture, Databases, Distributed Systems, among others. A second consideration is the demonstration that control theory can be used effectively by software practitioners in the development of commercial products, such as IBM's experience with autonomic features in DB2. The workshop seeks a unified view of the field that identifies commonalities between isolated research problems, provides opportunities to generalize, develops broad theoretical foundations for this new class of control problems in which the controlled process is a (typically distributed) software entity, and identifies likely application areas for practical impact. The workshop provides an opportunity for an interdisciplinary exchange to identify key research problems that are fundamental, cross-cutting, and most broadly applicable within the software system control domain. A compilation of the results is to be made publicly available and disseminated through published reports in control and computer science magazines and newsletters.