What Should Be the Size of a Source?

The size of a source may become a problem when it takes significant time to refresh or rebuild it. A source containing thousands of large documents can take hours to rebuild. Although the refreshing or rebuilding process does not prevent users from querying the source, it requires considerable CPU resources. Therefore, CES can be slowed down for several hours each day because of re-indexing. Creating smaller sources allows to distribute the refreshing or rebuilding process over the week or even the month.

Example: You can index the intranet of a company in three different sources—the first one can contain documents that are updated daily (ex.: project schedules), the second one documents updated once every few weeks (ex.: financial reports) and the third one documents rarely updated (ex.: documentation concerning previous versions of the product). Therefore, you can distribute the re-indexing process by refreshing the first source daily, the second one weekly and the third one monthly.

Note: The size of the index is limited by the license type. If a license allows the indexing of 100,000 documents, the total size of sources cannot exceed this number.