Authors:
Olga Ormandjieva
;
Heng Kuang
and
Stan Klasa
Affiliation:
Concordia University, Canada
Keyword(s):
Reactive Autonomic System, Performance Modeling, Self-monitoring, Category Theory, Representational Theory of Measurement, Decision Making.
Related
Ontology
Subjects/Areas/Topics:
Embedded Communications Systems
;
Software Architectures
;
Telecommunications
Abstract:
The research presented in this paper was motivated by the need to build performance self-monitoring and a decision-making process into Reactive Autonomic Systems (RAS). In order to achieve RAS compliance in terms of the imposed performance policies, we formalize RAS modeling and performance control in a single framework based on a representational theory of measurement and category theory. Category theory is expressive enough to capture qualitative and quantitative knowledge about heterogeneous RAS requirements and their interrelationships, as well as a decision-making mechanism, in one formal representation, where structure and reasoning are inextricably bound together. Thus, category theory provides a computational mechanism which enables this knowledge to be applied to performance data and RAS information structures in order to arrive at valid conclusions.