The general guidelines are as follows:

• 1 Collector = 10,000 Shares or 20 Filers, whichever comes first
• 1 Indexer = 20,000 Shares or 20 Collectors, whichever comes first
• 1 Management Server = 16 Indexers

• 100 MB per 1,000,000 files
• 20-30 MB per 1,000,000 events (read, write, etc.)

• 1,000 files per second, and 5-6,000,000 files per hour

 
Note: These are only initial guidelines for DI.  There are many factors that are involved in determining what storage/resources are required and it's possible some environments will need additional resources to allow Data Insight to run properly.