CAREER: An Infrastructure in Support of Configurable, Consistent, Interactive Computational Steering

9734103交互式转向允许用户以在线方式监控程序的执行并调整应用程序参数和资源分配。指导工具必须解决与监控、显示和交互组件的一致性、滞后、可伸缩性和诱导扰动相关的问题。然而，这些关注点之间的理想平衡因执行环境和目标应用程序而异。因此，可配置性是一个重要的问题。本研究的重点是设计和开发转向环境的基础设施，允许用户指定配置，以在一致性、滞后和扰动之间提供所需的平衡。该项目包括开发用于收集具有不同一致性程度的快照的模块化算法，以及允许将更改一致地应用于正在执行的程序的算法，同时最大限度地减少由此产生的延迟和扰动。这些用于一致性监视和引导的算法建立在现有快照算法工作的基础上，并通过乐观计算技术的创新应用，将这项工作从分布式系统的一致性观察领域扩展到确保对正在执行的分布式计算应用更改的一致性。这项工作的另一个有趣的特点是使用监测和指导“代理”，作为脚本实施，并部署在模块化、分级监测和指导环境中。*

9734103 Interactive steering permits users to monitor a program's execution and adjust application parameters and resource allocation in an online fashion. Steering tools must address issues related to the consistency, lag, scalability, and induced perturbation of monitoring, display, and interaction components. However, the ideal balance among these concerns varies with execution environment and target application. Thus, configurability is an important issue. This research focuses on the design and development of infrastructure for a steering environment that permits users to specify the configuration to provide the desired balance among consistency, lag, and perturbation. This project includes the development of modular algorithms for the collection of snapshots with varying degrees of consistency, and algorithms that permit the consistent application of changes to the executing program, while minimizing resulting lag and perturbation. These algorithms for consistent monitoring and steering build on existing work on snapshot algorithms, and extend this work from the realm of consistent observation of distributed systems to ensuring consistency in the application of changes to an executing distributed computation, through an innovative application of optimistic computing techniques. Another interesting feature of this work is the use of monitoring and steering ``agents'', implemented as scripts, and deployed in a modular, hierarchical monitoring and steering environment.***