I have been using a single MySQL database to drive an analytics dashboard that will will typically compute metrics and graphs based on ~1-10million records. The data set can grow quite heavily over a few days.
Needless to say, the next version of the dashboard will require approximately 100 time that amount of data, and I feel that MySQL will not be sufficient to report out these metrics in a very short period of time.
The new dashboard requirements look like..
- The ability to report metrics and graphs that will aggregate ~100 - 1000 million records
- The ability to report "some" basic metrics in real-time
- A very short period of wait-time for the end-user who is viewing the dashboard
After spending some time researching the different available database technologies, I have come to conclusion that the choice should be between one of the following:
NoSQL Databases
- Mongo DB
- Cassandra
Columnar Databases
- Infini DB
- Info-bright
Most of the information I found on the columnar database seems to be from the organizations that are actually working on developing those technologies. Therefore, I am looking for anyone who has an unbiased experiences with the columnar database and help me understand some of the challenges I will be facing if I go down that route. In addition, is a columnar database really something I should be using, or will one of the NoSQL database above sufficiently satisfy my requirements?
