Discussions specific to the VMware vSphere hypervisor
Post Reply
verre
Enthusiast
Posts: 67
Liked: never
Joined: Jun 30, 2014 1:51 pm
Full Name: Valentino Erre
Contact:

who makes Compression and deduplication?

Post by verre » Nov 24, 2014 4:33 pm

If I have a physical direct-san-access (fc) proxy, the deduplication and compression are performed by the proxy on the source datastore...is it true?
Therefore the data that pass between the VMFS and the proxy has already been deduplicated and compressed?
Thank you!

dellock6
Veeam Software
Posts: 5689
Liked: 1604 times
Joined: Jul 26, 2009 3:39 pm
Full Name: Luca Dell'Oca
Location: Varese, Italy
Contact:

Re: who makes Compression and deduplication?

Post by dellock6 » Nov 24, 2014 4:48 pm 1 person likes this post

correct, deduplication and compression are done by the source datamover, that is the proxy (at least for backups...).
Luca Dell'Oca
Principal EMEA Cloud Architect @ Veeam Software

@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2019
Veeam VMCE #1

Shestakov
Veeam Software
Posts: 6836
Liked: 697 times
Joined: May 21, 2014 11:03 am
Full Name: Nikita Shestakov
Location: Prague
Contact:

Re: who makes Compression and deduplication?

Post by Shestakov » Nov 24, 2014 4:49 pm

Hello Valentino,
verre wrote:the deduplication and compression are performed by the proxy on the source datastore...is it true?
Correct. Proxy performs on-the-fly deduplication and compression of data received from the source component.
verre wrote:Therefore the data that pass between the VMFS and the proxy has already been deduplicated and compressed?
Dataflow for backup job is: Disk > Backup proxy > Network > Backup repository > Disk. So data is deduplicated and compressed after backup proxy.
Tahnks.

alanbolte
Expert
Posts: 635
Liked: 172 times
Joined: Jun 18, 2012 8:58 pm
Full Name: Alan Bolte
Contact:

Re: who makes Compression and deduplication?

Post by alanbolte » Dec 01, 2014 4:06 am 3 people like this post

I saw this post in the Forum Digest and wanted to clarify - deduplication actually has to be done in two steps because you can have multiple source proxies. So Proxy A reads data blocks from VM1, compresses them, and deduplicates them against other data blocks from the same disk, and sends it to the target (repository) datamover. Proxy B does the same thing for VM2. The target datamover then deduplicates between the disks for both VM1 and VM2.

If your concern is the amount of data moving between the datastore and the source proxy, the main limitation is CBT. Its use with incremental backups is fairly obvious, but with active full backups on block storage it will read only the blocks containing data; it won't have to read any zero blocks. VMware has not implemented that feature of CBT for NFS storage.

Post Reply

Who is online

Users browsing this forum: Bing [Bot] and 10 guests