Proactive Computing Artificial Immune Systems Rogrio de Lemos

Proactive Computing: Artificial Immune Systems Rogério de Lemos University of Kent at Canterbury u Brian Bell (BAE Systems - Sowerby, UK) u Rogério de Lemos, Jon Timmis (University of Kent at Canterbury, UK) u Mark Neal (University of Wales, Aberystwyth, UK) u Andy Tyrrell (University of York, UK) Rogério de Lemos DEFINE – Pisa, November 2002 – 1

Biological Inspired Computing Autonomic computing (IBM) u nervous system which regulates the basic functions of the body; Planetary computing (HP) Self-healing systems (Software Engineering Community) Homeostasis (Mary Shaw) u a system acts to maintain a stable internal environment despite external variations; u react to change rather than desired states; Artificial immune systems u fault-tolerance, intrusion & virus detection; Rogério de Lemos DEFINE – Pisa, November 2002 – 2

Motivation Provision of means for a system to cope with changes: u design time (evolution): u u building new systems from existing ones; run time (adaptability); u adapting to changes that occur in the operating environment; Some approaches rely on learning capabilities/emergent behaviours: u adjust behaviour/structure to new needs without human intervention; Rogério de Lemos DEFINE – Pisa, November 2002 – 3

Artificial Immune Systems (AIS) AIS are adaptive systems inspired by theoretical immunology and observed immune functions, principles and models, which are applied to complex problem domains; What does the immune system offer? u pattern recognition; u robust and distributed; u adaptive and diverse; u learning and memory; Rogério de Lemos DEFINE – Pisa, November 2002 – 4

Proactive Computing Fault tolerance u Learning capabilities may enable the system to react to unexpected circumstances: u it removes the predictability aspect; System evaluation u how much can these learning capabilities be trusted? u how to protect the system from undesirable decisions? Rogério de Lemos DEFINE – Pisa, November 2002 – 5

Themes u Artificial Immune Systems applied to Fault and Intrusion Tolerance u u Biologically Inspired Engineering u u embryonics, evolvable hardware, immunotronics, etc. Biologically Inspired Dependable Systems u u e. g. , evolution of error detectors; Techniques for Autonomous exploitation of homeostasis Swarm Techniques applied to Intrusion Tolerance in Mobile Systems u Rogério de Lemos coordination of countermeasures to attacks; DEFINE – Pisa, November 2002 – 6