1. Using Store-procedure instead of SQL code in the program
Store procedures can improve performance. Many tasks are implemented as a series of SQL statements. COndirional logic applied to the resules of the first SQL statements determines which subsequent SQL statements are executed. If there SQL statements and conditional logic are written into a stored procedure, they become part of a single execution plan on the server. The results do not have to be returned to the client to have the conditional logic applied, all of the work is done on the server.
2. Normalized database, if not, the database will have lots of redundant data.
3. Over-normalize has no good for performance issue
If a database is over-normalized, i.e. the database is defined with numerous small, interrelated tables, database performance is reduced. This is because when teh database processes the data in numerous, small and interrelated tables, it has to combine the related data, which results in an increase in teh database's workload. In these situations, denormalizing the database slighly to simplify complex processes can improve performance.
4. Cursor
Cursor is efficient when the database is small, however, if the database's size increases dramatically, then using cursor could be a very time-consuming method compared to simple SQL statement.