Record storage is fairly easy to calculate, but index storage is not because index compression makes calculation difficult. The ever-evolving compression algorithms make the calculation even harder. You can run a database analysis and use the index-specific output to make your decisions. Remember to allow room for growth and general overhead, the same as with data storage.
If you have an existing database, you can take statistics to determine index storage size. Without a database, you have to estimate the size. The number and nature of indexes can vary greatly between applications. Word indexes and indexes on character fields tend to use more space, while numeric indexes are significantly more efficient in terms of storage. There are databases where indexes use more storage than data, but these are the exception and not the rule.
In general, indexes account for approximately 30 percent of total storage. Therefore, you can take 50 percent of your data storage as an estimate of index storage. Remember that this percent might vary greatly, depending on your schema definition. Consider this estimate as a starting point and adjust and monitor it accordingly.
The following example highlights a portion of a database analysis report that shows the proportion of data storage to index storage within an existing database. Use this information to determine the allocation of disk resources to the areas that are going to contain the data, as shown:
SUMMARY FOR AREA "Student Area": 8
----------------------------------
Records Indexes Combined
Name Size Tot percent Size Tot percent Size Tot percent
PUB.stuact 18.9M 12.6 9.7M 6.4 28.6M 19.0
PUB.student 30.3M 20.1 20.1M 13.4 50.5M 33.5
-----------------------------------------------------------
Total 115.3M 76.4 35.6M 23.6 150.8M 100.0
