SQL Server best practices (index optimization,Database Performance)

SQL Server 2005 Index Optimization Best Practices

Do you have a maintenance window to perform database reindexing?
Have you ever performed full set of re-organize and re-index process on a bigger database?
Do you know there are best practices to deploy in this regard?
On large database systems, with large numbers of insert and update commands, the problem of index fragmentation is one of the main causes of performance degradation and a proper index optimization strategy is a must. Also following are set of counters you need to keep in mind:
· Create Index on frequently used columns in T-SQL Code. Columns used in WHERE, ORDER BY and GROUP BY are good candidate for Indexes. Create Index on column which are used in JOIN Condition.
· Remove any un-necessary Indexes. As Index occupies hard drive space as well as it decreases performance of all the insert, updates, deletes to the table.
· Smaller Index Key gives better performance than Index key which covers large data or many columns
· Multiple Columns Index or Covered Index should be ordered as Most Selective column on left and gradually decreasing selectivity as they go right.
· Use SORT_IN_TEMPDB option when table is created if tempdb is on different disk. This will increase the performance to create Index.
Also there are few basic guidelines when you need to create a database from scratch such as,
· Design a normalized database.
· Optimize a database design by denormalizing.
· Optimize data storage.
· Manage concurrency - by selecting the appropriate transaction isolation level.
· Select a locking granularity level.
· Optimize and tune queries for performance.
· Optimize an indexing strategy.
· Decide when cursors are appropriate.
· Identify and resolve performance-limiting problems.
· Be familiar with index structures and index utilization. Specifically, they must understand the interaction between non-clustered indexes, clustered indexes and heaps. A must know why a covering index can improve performance.
· Be able to design a database to third normal form (3NF) and know the trade offs when backing out of the fully normalized design (denormalization) and designing for performance and business requirements in addition to being familiar with design models, such as Star and Snowflake schemas.

SQL SERVER - 15 Best Practices for Better Database Performance

I have written 14 best practices here, read them all and let me know what is as per your opinion should be the 15th best practice.
1. Store relevant and necessary information in the database instead of application structure or array.
2. Use normalized tables in the database. Small multiple tables are usually better than one large table.
3. If you use any enumerated field create look up for it in the database itself to maintain database integrity.
4. Keep primary key of lesser chars or integer. It is easier to process small width keys.
5. Store image paths or URLs in database instead of images. It has less overhead.
6. Use proper database types for the fields. If StartDate is database filed use datetime as datatypes instead of VARCHAR(20).
7. Specify column names instead of using * in SELECT statement.
8. Use LIKE clause properly. If you are looking for exact match use “=” instead.
9. Write SQL keyword in capital letters for readability purpose.
10. Using JOIN is better for performance then using sub queries or nested queries.
11. Use stored procedures. They are faster and help in maintainability as well security of the database.
12. Use comments for readability as well guidelines for next developer who comes to modify the same code. Proper documentation of application will also aid help too.
13. Proper indexing will improve the speed of operations in the database.
14. Make sure to test it any of the database programming as well administrative changes.