Limited Resources Platform Implementation of Complex Event Processing System

Abstract

Static although structured queries performed on data that is stored in relation databases or data warehouses are not sufficient enough for systems that are running in real-time, where the focus of monitoring is decided during the runtime of an application. Given that the cessation of activity of the core applications may not be admissible only compounded the problem. However, this solves complex event processing where time is perceived as more important than the actual parameter data persistence. There are currently a number of commercial and non-commercial CEP engines but most of them are very robust and non-optimized for use on platforms with limited resources. In this paper we propose modular and non-robust complex event processing solution for mobile devices with Android. This proposal is based on Publish Subscribe design pattern with a view to its possible further extension. Thus implemented prototype allows processing sensor data from various sources, both inside the device, as well as sensors from other connected devices. To demonstrate the usability of our approach, a mobile application for detection of potential threats to health was developed. This application, utilizing the CEP engine, can warn its users of any danger to their health, based on the predefined complex rules. We have analyzed several solutions and their application possibilities for use on platforms with limited resources. We have designed and implemented a prototype to support the development of complex events for the Android platform. Finally we have tested our approach with the existing solutions in comparable level.