10 million data non-real-time search, but requires rapid development, what advice?

Php+mysql+sphinx, the sooner the better, the faster the time.
Do you guys have any suggestions?

10 million data the table contains more than 30 fields, and there are not less than 10 child tables, and the fields that need to be searched are in both the main table and the child table.

The search is fast and the delay is guaranteed to be almost 1 seconds.

High concurrency odds are not too high

Reply content:

Php+mysql+sphinx, the sooner the better, the faster the time.
Do you guys have any suggestions?

10 million data the table contains more than 30 fields, and there are not less than 10 child tables, and the fields that need to be searched are in both the main table and the child table.

The search is fast and the delay is guaranteed to be almost 1 seconds.

High concurrency odds are not too high

The amount of data and the background you introduced, Sphinx should be competent?

1. Fully put in memory
2. Index optimization

Look at your description, my first reaction is. Unreasonable design ... Can you write down the detailed design or the demand, so that the plan ~ ~ ~
10 million data, Mysql+sphinx is absolutely capable, the design of a reasonable millisecond level is entirely possible.
Now I can think of:

• If necessary, sub-database or sub-table;
• Search all go sphinx,mysql only the key query;
• Optimize the table structure to reduce the probability of connection.

