I ran across this claim today – “ran across” means that Google decided that they should serve an advertisement to me while I was doing some web searches. I was interested (I’m the chief architect of Unitrends) so I clicked on over and found myself reading a bunch of material from a vendor who said they were “#1 Hyper-V Backup.”
The usual suspects were there with regard to features: near-continuous data protection (near-CDP), immediate failover without taxing the network, built-in compression and deduplication, and the like. (As an aside, Unitrends has all these features and a few more, and when we use terms like “deduplication” we’re not talking about simplistic job-based deduplication that doesn’t deduplicate across jobs like many vendors do.)
But none of it explained to me why the vendor was claiming “up to 20X faster.” I finally got to it – it was around CBT (Changed Block Tracking.) CBT is a technology which is readily available in VMware through the services they provide via VADP (vStorage API set for data protection); however, Microsoft doesn’t intrinsically offer these so vendors have to write software to offer them. Unitrends does, and these days most vendors do.
So why 20X faster Hyper-V backup? From what I can tell in the advertisement, they’re comparing themselves versus a full backup and/or versus a backup in which you have to scan the entire VM or something. But the sheer arbitrariness of 20X is what blew my mind – why not 100X or 1000X? I guess 20X just somehow sounds more credible – but if you’re throwing numbers around it’s pretty random.
So what do you think? What’s your greatest concerns with respect to Hyper-V backup? Raw performance? Recoverability? Ability to handle multiple hypervisors? Or ability to handle the enterprise – both virtual and physical?