A Generic Approach towards Measuring Level of Autonomicity in Adaptive Systems
International Journal on Advances in Intelligent Systems
This paper is concerned with setting the groundwork for the introduction of standards for Autonomic Computing, in terms of technologies and the composition of functionalities as well as validation methodologies. This is in line with addressing the lack of universal standards for autonomic (self-managing) systems and design methods used for them despite the increasingly pervasiveness of the technology. There are also significant limitations to the way in which these systems are assessed and
... e assessed and validated, with heavy reliance on traditional design-time techniques, despite the highly dynamic behaviour of these systems in dealing with run-time configuration changes and environmental and context changes. These limitations ultimately undermine the trustability of these systems and are barriers to eventual certification. We propose that the first vital step in this chain is to introduce robust techniques by which the systems can be described in universal language, starting with a description of, and means to measure the extent of autonomicity exhibited by a particular system. Existing techniques have mainly qualitatively classified autonomic systems according to some defined levels with no reference to the building blocks (core functionalities) of the systems. In this paper we present a novel and generic technique for measuring the Level of Autonomicity along several dimensions of autonomic system self-* (e.g., self-configuration, self-healing, self-optimisation and self-protection) functionalities. To demonstrate the feasibility and practicability of our approach, a case example of two different scenarios is examined. One example focuses on a specific case approach for LoA measure within a Dynamic Qualitative Sensor Selection scenario. The second example is a deployment of a generic case approach to an envisioned Autonomic Marketing System that has many dimensions of freedom and which is sensitive to a number of contextual volatility.