With the rapid pace of technological advancements in smart device, sensor and actuator technologies, the Internet of Things (IoT) domain has received significant attention. These advances have brought us closer to the ubiquitous computing vision. However, in order to fully realize this vision, devices and applications should rapidly adapt to the changes in the environment and other nearby devices. Most of the existing applications store collected data in a data store and allow users to query stored data to notice and react to such changes. Many of the applications utilize cloud and big data technologies for scalability. Nevertheless, the responsiveness of such IoT applications is largely limited due to polling based queries. Therefore, in this paper, we propose a centrally managed distributed infrastructure based on the state of the art big data technologies. We focus primarily on the problem of how to process multiple continuous queries in real time and notify users timely. Our main contribution is specifying a generic and scalable architecture to process a multitude of real time queries. We provide a prototype implementation to demonstrate the applicability of the approach. The scalability of the architecture is evaluated by conducting several experiments on the prototype implementation.