Metabolic Computing

In this paper, the author proposes a metabolic computing model for a truly renewable system with high fault tolerance and sustainability and a realistic architecture for the model using four kinds of elements: metaboloids, slots, a power queue, and recycle unit. Metaboloids, which are processing units, are arranged in a mesh in the power queue. However, as the metabolism may change the network, to manage running tasks, metaboloids must achieve homeostasis, for which two new algorithms, bubbling and drifting, are presented. For simple metabolism, the specification of the architecture does not change. Consequently, despite using innovative hardware manufacturing technologies, performance does not improve. The author proposes an evolutional architecture for a metabolic computing model, whereby a set of metaboloids with different specifications can work together if the specification is portable and the difference between specifications is at most 1. Finally, the author presents a reflective architecture for the power queue, minimizing the number of unrenewable units.

The Timeless Way of Building REA Enterprise Systems

In a continuously changing business environment, the need for enterprise systems that are more adaptable has been recognized by many. Several solutions are being suggested to improve the adaptability of enterprise systems, including service-oriented architectures, model-driven architectures, and reflective architectures. In this paper, we propose a timeless way of building enterprise systems that employs a reflective architecture with integrated Resource-Event-Agent (REA) enterprise ontology specifications. We show how the explicit recording of enterprise schema descriptions results in enterprise systems with increased adaptability. In addition, we demonstrate how explicitly recorded ontological specifications can further increase application reusability. We validate our research with a prototype system.