With the increasing complexity of computing systems follow dramatically increasing management problems and costs. IBM's Autonomic Computing initiative proposes to reduce complexity by making computing systems more self-managing: self-configuring, self-healing, self-optimizing and self-protecting, from individual components to complete environments [1].

An increasingly popular approach is to apply concepts and techniques from automatic control and related fields. A simple illustrational example is managing the CPU load and memory use of a web server using traditional linear control [2].

Controlling a networked system is much harder. The system components all have local observation and control points and incomplete knowledge of overall system state. The effects of controls are uncertain, interact, and may appear long after their application. For such problems, recent advances in approximate dynamic programming - a field at the intersection of control, operations research, and computer science - offer us new powerful tools.

In this VINNOVA-funded feasibility study, we explore self-managment methods for networked systems, such as ERP systems, telecom systems and command and control systems.

Researchers

• Sverker Janson, Ph.D.
• Joakim Eriksson, M.Sc. (Ph.D. student)
• Niclas Finne, M.Sc. (Ph.D. student)
• Patrick Hinnelund, Master student (KTH)

Collaboration

• IFS AB, Pär Hammarström

Project Results

• Article in ErcimNews: Towards Self-Managing Systems
• Master's Thesis: Autonomic Computing - a method for automated systems management
• Presentation: Autonomic Computing (held at IFS)
• Poster: Selfmanaging sensornets (shown including demo at SICS open-house and SITI conference 2004)

Links

• IBM Autonomic Computing initiative

References

1. J. Kephart, D. Chess. The vision of Autonomic Computing. IEEE Computer, 2003.
2. Y Diao, N Gandhi, JL Hellerstein, S Parekh, and DM Tilbury. MIMO Control of an Apache Web Server: Modeling and Controller Design. American Control Conference, 2002.