Replication speed

[mchoiniere]

Hello,

I was wondering what was faster, replication or backup. I supposed that backup was slower since it has to compress/dedup the data before sending it while the replication just sends the data. Also, on big vmdk ( > 1TB), the replicas are very long, mainly because of the snapshot deletion. Is there something to do for that ?

Thanks.

Marco.

[foggy]

Marco, replication jobs also have compression settings (for optimal traffic transfer between source and target). So, with all the rest conditions identical, the job duration will mostly be defined by the backup method (forward/forever forward/reverse) and target storage/datastore performance. Also, replica snapshot deletion (when retention comes into play) can indeed affect the entire job duration.

But why the question? Backup and replication serve different needs and should be both used utilized to achieve better protection.

[mchoiniere]

We are using both right now. The problem I have is that my DR setup has a slower san and we have large vmdk ( > 1TB) and the snapshot deletion takes longer than the replication itself. I compared with the backups on our faster san and it's much slower. I was wondering if I was missing something. I wonder why snapshot deletion is so long.

Marco

[dellock6]

You mean the snapshot on the souce VM, or the snapshot at DR that is removed when retention is reached? If the latter, as you say there's a slower SAN it may be possible that the deletion of the oldest snapshot is taking a while, because technically you are not deleting it, rather injecting its data into the parent VM.

Luca