Foundations of Autonomic Computing Development

Abstract

The complexity of modern middleware, and software solutions, is growing at an exponential rate. Only self-managing, or autonomic computing technology can reasonably stem the confusion this complexity brings to bear on human administrators. While much has been published on "architecture" and "function" for producing such systems, little has been written about the engineering of self-managing systems as a distinct paradigm. In this paper we suggest a straw-man for engineering of autonomic systems that is based on two essential tracks: a set of engineering principles that should guide the planning of autonomic systems and their interfaces and secondly a set of mathematical foundations upon which such systems can best be constructed. These foundational attributes are intended to guide the thinking of R&D organizations pursuing the development of autonomic computing capability. The role of architecture and standards is also discussed, highlighting their role in inter-component management.

Targeting Prediction: Engineering a Distributed Event Processor for an Autonomic Biometric System

“…Currently within the Autonomic Computing community much of the policy research is a misinterpreted that equates to policies = rules; or that there is a very close connection [12] yet when you consider for example 10% growth in sales as a business policy/target as opposed to a rule concerning a parameter, the need for prediction at a systematic higher level becomes clear.…”

Role-based systems are autonomic

“…(As You Like It, Act II, Scene 7)". Role-based approaches are good candidates to build autonomic computing systems, because "autonomic computing technology must work in real world scenarios [12]." Definition 6: workable role.…”

“…Assignment of potential agents to roles is a fundamental for dynamic optimization in role-based systems. The following rules (10)(11)(12) are used to help a role-based system do self-optimization.…”

A Quality-Driven Approach to Enable Decision-Making in Self-Adaptive Software

“…Lightstone discusses a few of the foundation techniques for solving self-management problems, such as expert system, tradeoff elimination, static and online optimization, control theory, and correlation modeling [119]. Each of these techniques may define entities in each of the three spaces differently and compose different structures.…”