Relational databases are ideal for many applications, but not the best solution for timeseries data.
Historian applications have been around a long time, long enough so that certain preconceived notions, unfounded impressions, or ‘myths’ have developed through the years that seem to perpetuate without substance or evidence. If you want to better understand the merits and shortfalls of relational databases as well as uncover and challenge these myths in the context of process historian applications, then read on. It will also aid in understanding the basics of historian applications, discover what to consider during an evaluation, and properly set your expectations on the value a historian can bring.
Just because the New York Stock Exchange trading systems and other high-throughput applications use a relational database doesn’t mean they’re right for everything. Take time-series data, for example: can SQL Server or Oracle store time-series data? Certainly. Are there some problems you should know about before you try it? Absolutely