Table keys are used to identify records, and to speed up filtering and sorting. Having too few keys may result in painfully slow inquiries and reports. However, each key incurs a processing cost, because the index containing the key must be updated every time information in a key field changes. Key cost is measured primarily in terms of increased index maintenance processing. There is also additional cost in terms of disk storage space (usually not significant) and additional backup/recovery time for the increased database size (sometimes very important).
When a system is optimized for processing speed, it is critical to analyze the SQL Server indexes that are active because that is where the updating and retrieval time are determined. ...
