Speeding up replication...

You can of course do the same thing when performing an initial back up of a large remote client

Exactly right.  This is a recommendation which we make when a customer has a large amount of initial data to replicate and a limited amount of bandwidth.  It's sometimes referred to as 'seeding a replication target'.

Thank you very much :-)

Yes, Avamar does the de-duplication at the source, it first scans the files and check the local cache if that file is been backed up or not else ask the server if it has the copy of "content" of the file, if server says yes it will skip and proceed with the next file and update the local cache.

It does not matter which file resides on which server in your environment, Once the file is backed up on avamar Server it will never backp again

Regards

Amol Powar

I found this interesting.  So it doesn't matter of which client or location or physical drrive teh file is located?

If I create a PowerPoint presentation then save it and back it up.  Then email that PP to another user who opens it and saves it to their desktop.  When avamar backs up their PP, it will only backup the changes even though it is on another machine, in another location, under another user?  If so... .that is AMAZING!!

Does it check the local server based on a hash or file name?

First it generates the hash with the file name and checks the local cache and then on server(if not present in local cache), if reference is not found on both the location it will chunk the content of the file and generates hashes for each chunk and repeat the process