Why industrial manufactoring data need special considerations
In respect to storage, retrieval, aggregation and reporting of industrial manufacturing data, special considerations are needed. Data are immutual and written only once. Customers are normally interested in streams and aggregations, not in a single data record. Really "Big Data" may apply.
In one of our projects, we have to store more than 3 million measurements every second into PostgreSQL, resulting in a Terabyte database. Technicians expect to evaluate both live data and historical data within an acceptable period of time. Not only within the company's network, but also outside using tablet or smartphone.
By example of real live industrial projects, the requirements and technical aspects of our solutions will be explained. This includes how the required performance is achieved by using PostgreSQL array features and XML capabilities.