Nowadays, the overwhelming amount of data generated on Internet have boosted the development of new solutions to process data on time windows closer to real time. Such solutions are known as Distributed Stream Processing Engines (DSPEs). DSPEs were specially designed to process data streams over cluster infrastructure, however the great massification of mobile devices opens new opportunities to process data closer to the source in order to reduce latency and traffic over the network. In this paper we propose Symbiosis, an architecture oriented to process data streams over mobile clients such as tablets and smartphones. Symbiosis model aims to exploit mobile devices resources to pre-process data streams generated on neighbors clients. Implementing data processing on mobile nodes is challenging due to their mobility and the limited battery power. In order to cope with such a requirements, Symbiosis proposes a data processing method based on checkpoints which consider both mobility and available energy in the device.

Every year, thousands of disasters impact the world, generating substantial loss in terms of lives and public infrastructure. The recovery phase requires substantial money investment for governments, in order to rebuild the affected region. To minimize the effects of a disaster, applications supporting decision-making through real-time analytics must be fast and accurate. These tools may be supported by stream processing systems that enable the real-time processing of structure-less data; however, their use and configuration require expert programming knowledge and a setup phase that could lead to loss relevant data. This work proposes a platform based on stream processing systems for supporting the analysis of large streams of data emitted in case the event of disasters. This platform facilitates deploying applications tailored for specific disaster scenarios, provides elasticity in case of changes in the amount of traffic and enables applying migration to increase robustness. We have built a study case to detect needs in case of an earthquake and discuss the properties of the system.