Microsoft has formally launched Azure Site Recovery, its service that allows one to fail over from an on-premises bit barn to an Azure facility.
Redmond says that acquisition's technology has been thrown into the Site Recovery mix and the service can now handle “non-virtualized physical servers and heterogeneous hypervisors”.
The service operates a bit like Dropbox for VMs: users can set a policy to determine which virtual machines are mirrored in Azure, complete with schedules for snapshots and data synchronisation uploads. VMs are then uploaded to Azure on the specified schedule and, if one's data centre suffers a misfortune, the VMs will start to operate in their Azure incarnations instead. System Center integration means there shouldn't be too many new bits to wrap one's head around, although there is agent software to install.
Just what recovers to where and when is determined by creating recovery plans using a dedicated portal. That facility allows recovery to another data centre or to Azure. The choice is yours, but the recovery service runs in Azure.
Pricing starts at US$54 per instance, per month. Users also need to set up Azure storage, which will incur extra charges.
Microsoft's pitching the service as cheaper than building and operating a DR site. A hypothetical rig with 100 VMs would cost $64,800 a year, plus storage. Even if the storage costs another $64k, the $128k total is a less than many midsize arrays. And that's before one considers operational costs for a DR site.
Azure's not alone in offering such services: VMware's vCloud Air has a comparable DR offering, with the same One Control Freak To Rule Them All ideology.
With the numbers stacking up so sweetly, it's hard not to imagine this kind of thing will soon become a common way to do disaster recovery. ®