Question

I’m looking to design a system that has the following data capabilities:

1. Multiple sources high writing rate, 50K – 100K of records per second (from ).
2. Single source fast reading, single IO fetch operation.

This is basic DB table/set design:

  Type name
--------------
Date
Amount

• For every new Type a new table/set will be created.
• The only fetch operation is SELECT by dates (should be clustered primary key or something similar).
• Older than one month tables will be archived.

The offered solutions I’ve received so far are:

• Pure DB engine such as Berkeley or MS Jet.
• NoSQL such as Mongo DB.

P.S – the system will be written in C#.

Answer 1

Redis would be the perfect solution, but the requirements are much simpler than that and it appears that Berkeley DB would suit just fine.