Which process divides the data into chunks for EMC Avamar deduplication data flow processing?

Explanation/Reference:

Assume the file HAS been changed, and therefore the hash of the file no longer matches the hash in the filecache. From here, the data will be broken into variable segment lengths through Avamar’s patented sticky bytefactoring process. These segments are then compressed andhashed, and the resulting hash is written to aHASH CACHE file locally on the client. This file is also loaded into RAM when a backup is initiated.

http://www.egroup-us.com/2011/09/emc-avamar-how-source-global-dedupe-magic-happens/

Avamar, has a rather nifty technology called Sticky Byte Factoring which allows it to identify the changedinformation inside a file by breaking the file into variable length objects, this leads to much greater efficiencythan fixed size approaches as changes made early in a fixed length sequence affect all subsequent blocks/chunks/objects/whatever in that sequence. This in turn changes all the fingerprints taken following the changeddata which means you end up with a lot of new blocks/chunks/objects/whatever even if the data really hasn'tchanged all that much. Sticky Byte Factoring on the other hand can tell what exactly has changed, not just thatthings have changed.

http://storagezilla.typepad.com/storagezilla/2007/09/the-future-is-v.html