Your dilema is exactly what Nimble is trying to avoid, complex backup software where it is not always needed. I've very recently come across a recovery solution that is really interesting - please read the rest:
First, a Nimble Array with snapshots at the disk/lun level (Redirect On Write (ROW) snapshots - i.e. no impact when taken and stored efficiently (with real-time compression before hitting disk)) with built in Replication (SAN to SAN) capabilities at no extra cost.
The "backup" part is then covered by snapshots and for extra protection you can ensure these are VMware (vCenter) or VSS consistent snapshots (as appropriate) or perhaps for those large file-system shares, just a crash-consistent snap (We all know MS survives these quite happily!)
So protecting the data is easy, recovering it not always as easy.
I've just come across a bit of software out there called "UFS Explorer" (no tie to Nimble in any way that I'm aware), what this allows you to do is mount a snapshot or clone a snapshot (from any snapshot taken on the array) and mount this to any host (Windows/Linux/MAC) that can connect to the array via the iSCSI protocol, then you browse the content (LUN/Volume) and recover individulal files (makes sense with NTFS/EXt... etc).
Now what impressed me even more is the ability to mount a VMFS filesystem (DataStore), browse to an individual VMDK file and then browse the content of the VMDK file at the file level.
So now you can take snapshots hourly (or more regularly if needed) and when a file recovery is needed on a VM, simply connect from your workstation to the relevant snapshot, browse and recover!
Unfortunately Hyper-V virtual disks is a "future"
OK, there is no index/catalog but how oftern do you really need to search an index/catalog? You know what it is you want to recover and your Recovery Time Objective (RTO) has now improved from daily (at best with backup software) to hourly or less.
