Many Smart City applications are designed following event-driven model, which means they react to new events and context changes. The volume and variety of data are important aspects of Big Data processing with new challenges added by Smart City applications. In particular,

Smart Transportation connects various types of data that are collected with a high velocity (toll booths, traffic patterns, meters, eye witnesses, video cameras) in order to use them efficiently and avoid traffic jams in the context of continuously growing of the population. Water management detects the pipes which are close to break and generate alerts for skilled crew to fix the pipe before this pipe breaks, predict water demand and help setting price schedules. Patterns can be detected and seasonal peaks can be forecast. These examples require real-time processing and extraction of valuable data


Smart Cities applications have specific data access patterns (frequent, periodic or ad-hoc access, inter-related data access, etc.) and address specific QoS requirements to data storage and processing services (response time, interrogation rate, etc.) With the advent of mobile devices (such as smartphones and tablets) that contain various types of sensors (like GPS, compass, microphone, camera, proximity sensors, etc.), the shape of context-aware (or pervasive) systems changed. Previously, context was only collected from static sensor networks, where each sensor had a well-defined purpose and the format of the data returned was well known in advance and could not change, regardless of any factors. Mobile devices are equipped with multimodal sensing capabilities, and the sensor networks have a much more dynamic behavior due to the high levels of mobility and heterogeneity. Aside from the benefits brought by using mobile devices to gather context information, as having significantly more computation, communication and storage resources than mote-class sensors, new challenges appear.

These challenges generally concern the way context information is gathered, analyzed, represented and stored, which is related to Big Data Analytics. Our proposal will, in this sense, provide answers to most of these challenges, as it will offer a platform designed for the storage of context data in the cloud. In this sense, we assume a generic framework, where context information can be acquired either through direct sensor access, or by employing a middleware infrastructure or a context server.