A lot of IoT data is basically:
time → measurement
For example:
10:00 → 68.2
10:01 → 68.5
10:02 → 69.1
10:03 → 69.0
It looks simple.
Then you multiply that by thousands of devices sending readings all day.
Now you've got a database problem.
You have to think about:
How much data are we storing?
How quickly are we writing?
What queries will we run?
How long do we keep raw readings?
Do we really need every measurement forever?
One thing I find interesting is data aggregation.
Maybe recent data stays at one-minute resolution.
Older data becomes hourly averages.Even older data becomes daily summaries.You don't necessarily need to keep every raw measurement forever.The right storage strategy depends on what you actually need to ask the data later.That's probably a more important question than simply asking:
"Which database should I use?"
Top comments (0)