JoWUA

Abstract

The number of users of handheld devices will exceed one billion in the coming five years1. These devices are increasingly being enhanced with new sensors, which enable the development of contextaware mobile applications. Moreover, mobile applications might share the same contextual information (decision logics) which in return shares data from the same sensors; this introduces high code redundancy. Scalable extensible middleware framework for context-aware mobile applications (SCAMMP) simplifies the aggregation and sharing of raw data from different sensors and the dynamic injection of decision logics in order to generate high level contextual information. It allows mobile applications to share these high level contextual information (decision logic) via a simple Application Programming Interface (API). This paper presents a fully-implemented SCAMMP (on Android platform) with a quantitative performance analysis. The analysis shows that SCAMMP’s overhead due to power consumption is around 0%, processing power have peaks less than 14% at certain moments but is zero most of the time, and the memory usage did not exceed 5 MB. It also shows that SCAMMP maintains its scalability after injecting additional decision logics and incorporating more sensors. Furthermore, SCAMMP allows applications to access high-level contextual information by adding only three lines of codes.