I have a scenario where in the table called USER_RESPONSE_SUMMARY there are 20 records and 2500 concurrent users/threads are
hitting the db and trying to update those 20 records.No new inserts.
Due to this we were having locking problems with the table.
This table just tracks the user response summary in aggregation where as actual USER_RESPONSE table stores responses.

I am contemplating the following 2 options.
1. Use materialized view
2. Batch the responses with some configuration to aggregates the counts and then update the table with counts using native query.
3. Keep inserting into the table with latest batch counts.

I'm wondering Which one is the better approach? any suggestions?
How to handle the concurrent requests(2500) more than table(20 rows) can handle?

Thanks
Krishna