An Adaptation Reasoning Approach for Large Scale Component-based Applications

Abstract

There is a growing demand for context-aware applications that can dynamically adapt to their run-time environment. An application offers a collection of functionalities that can be realized through a composition of software components and/or services that are made available at runtime. With the availability of alternative variants of such components and/or services that provide the basic functionalities, while differ in extra-functional characteristics, characterized by quality of services (QoS), an unforeseen number of application variants can be created. The variant that best fits the current context is selected through adaptation reasoning, which can suffer from the processing capabilities of resource-scarce mobile devices, especially when a huge number of application variants needs to be reason about. In this paper, we present a reasoning approach, which provides a meaningful adaptation decision for adaptive applications having a large number of variants within a reasonable time frame. The approach is validated through two arbitrary applications with large number of variants.

Keywords: self-adaptation, ubiquitous computing, adaptation reasoning, variability, scalability, utility function