Basically come with an sql table that contain a million rows. Let us say a person table.

Which kind of software will i need, to be able to handle 10 read/write every second. I believed of utilizing a Java NIO server to handle connections.

But exactly how does the rear-finish Database work? Could I merely use MySQL on a single computer?

Any insight could be great. Links, reading through, good examples. books?

I understand SQL. I've done a lot of SQLite but never produced a scalable system additional type of load.

Edit update,regarding helios comment

the number of reads versus. creates?: 50/50
do you want up-to-date-reads(no delay): YES?
how large is the items?: 10% is 10-15 posts and also the relaxation is 1-3 posts
are you currently being able to access them individually?: NO, non from the USER threads are interacting but there might be synchronised DB read/write on same row, (simply make it synchronized?)

Should you take advantage of JDBC Connection Pooling (like C3PO, DBCP etc.) you'd have the ability to have parallel card inserts, and also you would have the ability to have 10 threads (or even more) concurrently placing data. Your limit would then become your platform assets (memory, I/O etc.).

All of this would hold however only when the information insertion process itself could be parallel threads (i.e. you don't have a particular requirement to place records sequentially) which your work are pretty straight forward card inserts and never something complex that locks the table or causes another transactions to hold back.

Also think about using JDBC prepared claims, as well as carrying out in batches instead of after each record. This could accelerate things greatly.

which means you need 10 transcation/second on table with million rows.

that's really neither huge data set nor high end.

MYSQL (presently 5.5+ ,innodb engine) , running on single server,can certainly handle that.

you might need read first five chapter of 'High Performance MySQL' released by oreilly.

for nosql-db, i would recommend mongodb, see