Grand Challenge: Experiment-Driven Adaptive Systems
Source: Duke University
The collection, maintenance, and use of instrumentation data from systems is a central theme in autonomic computing research. Instrumentation data, collected through various sensors, can be used in many ways to simplify system management. For example, the data can be used to understand and forecast complex system behavior, learn models of system behavior to aid failure diagnosis and capacity planning, and derive controllers that automatically tune system configuration under varying workloads and recover systems upon failure.