Title: 
Model Based Analysis and Design of Self-Managing Computing Systems

Abstract
-------
Modern computing systems support a range of mission-critical
information technology applications crucial to commerce and banking,
transportation, and command and control systems, to name just a
few. Consequently, their proper design and operation have significant
economic and social impact. To operate such systems effectively while
maintaining the desired performance multiple operational parameters
must be dynamically tuned to adapt to changing state, requirements,
and operating conditions. As system and application scales increase,
ad hoc heuristic-based approaches to application adaptation and
management quickly become ineffective. Model-based technologies help
address this problem by enabling design-time analysis and providing
means to automate the development, deployment, configuration, and
integration of computing systems.  This presentation introduces recent
work on developing model-based approaches for systematic design of
self-managing computing systems. The developed approaches use
mathematical models to represent the system reaction to both control
and environment inputs. In these approaches, management problems of
interest are posed as a sequential and discrete optimization under
uncertainty. Results of this work show that model-based techniques can
be effectively applied to predict, analyze, and manage the complex
event-driven behavior of modern computing systems, under normal and
abnormal operating conditions. The presentation introduces several
implementations of this model-based technology and discusses future
related research directions.