Defining Complexity Factors for the Architecture Evaluation Framework

Abstract

The design and implementation of telecommunication systems is an incremental and iterative process, and system architectures may need to be revised and refined several times during their lifetime. Formal evaluation facilitates the identification of the weak points, where improvements are due in these architectures. In the domain of telecommunications, such evaluation can be based on the Architecture Evaluation Framework (AEF). During the evaluation, a deep understanding of the processes within a system is needed. Meanwhile, the systems being designed are usually complex systems encompassing a large number of components with an intricate pattern of interaction between them. As a result, it is extremely difficult to understand, predict and control the behavior of such systems. Theoretical studies in the field of complex systems describe potential reasons of system complexity, and explain its possible outcomes, as reflected in system structure and behavior. This knowledge may be utilized in architecture evaluation, in order to deepen the understanding of the interactions imposed by the architecture, as well as to extend the understanding of the involved architectural tradeoffs. For this, the complexity factors should be taken into account during the evaluation. However, no such factors are involved in the current version of the AEF. In this paper, the attempt is made to identify how the knowledge about properties of complex systems could be utilized for the evaluation of information system architectures. Based on the theoretical advances in the field of complex systems, a list of the complexity factors to be included in the AEF is compiled. These factors are going to be further refined, as the AEF is employed for evaluating real-world architectures.