Advice for a simple in-memory database server (persistence not required)
Multiple connections should be supported, preferably via ODBC. The client will run as a separate process on the same machine. Persistence is not required because the client will handle persistence elsewhere. If important, the client is written in C++.
The data is simple, a set of unrelated two-way graphs. Accessed directly through a value or range (
between X and Y), no updates. We don’t actually need SQL here, so non-SQL solutions can also be considered.
Client applications are multi-process and can run on multiple machines. Each machine should have a local copy of such a database, which is updated by its local client against the central storage.
- The platform is Linux
- For security reasons, RAM disks are not an option – we don’t want anyone with access to the machine to be able to view the data
- Data can only be saved in encrypted form, so the solution either doesn’t save the data at all or allows it to be saved using user-defined filters/plugins.