10 million non-real-time data search, but fast development is required. What are the suggestions?

Php + mysql + sphinx, the faster the time, the better. Do you have any suggestions? 10 million the data table contains more than 30 fields, and there are no less than 10 sub-tables, fields to be searched are involved in both the master table and the sub-table. The search speed is required, and the latency is guaranteed to be about high concurrency within 1 second... php + mysql + sphinx. The faster the time, the better.
Do comrades have any suggestions?

10 million the data table contains more than 30 fields, and there are no less than 10 sub-tables. The fields to be searched are involved in both the master table and the sub-table.

Search is required to be fast, and the latency is almost within 1 second.

High concurrency probability is not high

Reply content:

Php + mysql + sphinx. The faster the time, the better.
Do comrades have any suggestions?

10 million the data table contains more than 30 fields, and there are no less than 10 sub-tables. The fields to be searched are involved in both the master table and the sub-table.

Search is required to be fast, and the latency is almost within 1 second.

High concurrency probability is not high

In the context of your introduction, should sphenders be competent?

1. Fully stored in memory
2. INDEX OPTIMIZATION

According to your description, my first response is .. Unreasonable design... Can you write down the detailed design or requirements ~~~
10 million of the data, MySQL + sphinx is absolutely competent, and the design is reasonable within milliseconds is completely possible.
What I can think of now:

• Database or table sharding if necessary;
• All searches go through Sphinx, and MySQL only performs primary key queries;
• Optimize the table structure to reduce the connection probability.