I've lately been researching NoSql options. My scenario is the following:
We collect and store data from custom hardware at remote locations all over the world. We record data of the many site every fifteen minutes. We'd eventually prefer to proceed to every one minute. Each record has between 20 and 200 dimensions. Once setup the hardware records and reviews exactly the same dimensions each time.
The greatest problem we're facing is the fact that we obtain another group of dimensions of the many project. We measure about 50-100 different measurement types, however any project might have a variety of each kind of measurement. There's no preset group of posts that may accommodate the information. Due to this we create and make each projects data table using the exact posts it requires once we setup and configure the project around the system.
We offer tools to assist evaluate the information. This typically includes more information and data aggregation, most of which we store.
We're presently utilizing a mysql database having a table for every client. You will find no relations between tables.
NoSql appears promising because we're able to store a task_id, timestamp then your relaxation wouldn't be preset. What this means is one table, more associations within the data, but still handling all of the dimensions.
Is really a 'NoSql' solution suitable for this task? If that's the case which of them?
I've been analysis MongoDB also it appears promising...
Example for Clarification:
Project 1 has 5 data points recorded, the mysql table posts seem like: timestamp, temp, wind speed, precipitation, irradiance, wind direction
Project 2 has 3 data points recorded mysql table posts: timestamp, temp, irradiance, temp2