IoT Big Data Management for Improved Response Time

Internet of Things (IoT) impact in our daily life quickly evolves from recording basic data to process and take decisions based on complex data. Data provided by various sources (Health Care IoT, life quality, smart cities etc) came in many formats and will be used for statistical analysis and decision making, often based on data evolution and machine learning. Decisions involving data analysis over a huge database can be very time consuming. In critical time situations, system response time needs to be highly improved. This paper discusses a proposed architecture, suitable to improve the response time on systems using data analysis and decision based on data evolution. Simulation results for a system using sensors, gateways and persistence layers are presented and discussed. Considering all sensor data is sent to a non-relational database (NoSQL) suitable to store various data formats, this paper discusses the particularities of using IoT programmable gateways to send data used in critical time analysis to a fast relational database (SQL) database. In such databases, queries are very fast and, as a result, the decisions based on data evolution have an improved time response. Analysis to detect critical data patterns is triggered when data is inserted in SQL database or based on a time interval. Both strategies are discussed and analysed.