Rule-based monitoring and steering of distributed scientific applications

Abstract

This paper presents the design, prototype implementation and experimental evaluation of DIOS++, an infrastructure for enabling rule based management and control of distributed scientific applications. DIOS++ provides: 1) abstractions for enhancing existing application objects with sensors and actuators for runtime interrogation and control, access policies to control accesses to sensors/actuators and rule interfaces, and rule agents to enable autonomic monitoring and steering; 2) a hierarchical control network that connects and manages the distributed sensors and actuators, enables external discovery, interrogation, monitoring and manipulation of these objects at runtime, and facilitates dynamic and secure definition, modification, deletion and execution of rules for autonomic application management and control. The framework is currently being used to enable autonomic monitoring and control of a wide range of scientific applications including oil reservoir, compressible turbulence and numerical relativity simulations.