Article ID Journal Published Year Pages File Type
459800 Journal of Systems and Software 2013 17 Pages PDF
Abstract

Software systems that provide context-awareness related functions in pervasive computing environments are gaining momentum due to emerging applications, architectures and business models. In most context-aware systems, a central broker performs the functions of context acquisition, processing, reasoning and provisioning to facilitate context-consuming applications, but demonstrations of such prototypical systems are limited to small, focussed domains. In order to develop modern context-aware systems that are capable of accommodating emerging pervasive/ubiquitous computing scenarios, are easily manageable, administratively and geographically scalable, it is desirable to have multiple brokers in the system divided into administrative, network, geographic, contextual or load based domains. Context providers and consumers may be configured to interact only with their nearest, relevant or most convenient broker. This setup demands inter-broker federation so that providers and consumers attached to different brokers can interact seamlessly, but such a federation has not been proposed for context-aware systems. This article analyses the limiting factors in existing context-aware systems, postulates the design and functional requirements that modern context-aware systems need to accommodate, and presents a federated broker based architecture for provisioning of contextual information over large geographical and network spans.

► A novel federated broker model for provisioning of contextual information in large-scale distributed context-aware systems. ► A Mobile Context Broker that participate in the broker federation to facilitate smartphone based context applications. ► Empirical evaluation of load scalability of the federated broker model, in comparison to the centralised model. ► Energy conservation benefits of the Mobile Context Broker on smart phones during context provisioning and utilisation. ► Evaluation and analysis based on practical experiments with empirical results.

Related Topics
Physical Sciences and Engineering Computer Science Computer Networks and Communications
Authors
, , , , ,