Virtualisation held the promise of simplifying IT infrastructures, but ended up creating another abstracted management layer, says Veeam’s CEO, Ratmir Timashev
“I see virtualisation as a disruptive technology that will change the way people do IT and computing, in particular systems management, disaster recovery and business continuity.”
That’s the upside. But there’s a downside, says Ratmir Timashev, chief executive of Veeam – and naturally enough, that problem is what he’s dead keen to help us all with.
Most management tools were designed for the physical world, he says, and don’t cut it in the new world: “The virtual world is very different and those tools for the physical tools don’t work well in the virtual space. I believe we have the opportunity to be the largest systems management vendor in this [virtual] space,” Timashev said.
Veeam’s roots go back to Aelita, a Windows Server, Active Directory and Exchange management company: “We sold that part of the business to Qwest software in 2003. The same team founded Veeam in 2006 to fill gaps and holes we saw within enterprise needs in the virtual space, focusing first on the market leader, VMware (read our review of VMware’s vSPhere 4). But we are also planning to extend the functionality of our products to Microsoft’s Hyper-V.”
Virtualisation eases recovry, makes back-up harder
“One major issue is virtual disaster recovery,” he said. “Virtual provisioning has created new ways to make more resources available than in the physical world, with faster recovery. So there are new opportunities for disaster recovery, but they also create new challenges around performance management, capacity planning, reporting, chargebacks, auditing and VM [virtual machine] sprawl.” These issues don’t exist in the physical world, he said.
“As it is self-contained, a VM takes time to recover. But it’s also much easier to recover a VM than an entire server, the operating system, its applications and so on from scratch. So, a disaster that would have traditionally taken days to recover from, now takes minutes.”
Recovery can be easier, but back-up can be harder: “In the physical world if you want to back up a hundred servers, you would send an agent to each server and back them up in parallel. But in the virtual world this doesn’t work because all the VMs write to one one host and therefore take up CPU, memory and other compute resources.
Although VMware offers consolidated back-up, it’s combined with shared storage and snapshot technology, which creates issues: “Often the VM is trying to backupm and continues to backup the snapshot – not the VM itself,” said Timashev. “So you have to run the VM and the host in parallel.”
Virtualisation also affects pricing for software: “Symantec has a 60-70 per cent share of the Windows Server market, but it licences its back-up product per physical server.”
This can make Symantec expensive: “The same pricing system in the virtual world would apply to each VM. This adds up to a regular retail price per host of around $3,000 (£1,856). Whereas we charge $500 (£309) per CPU socket. VMware can cost as much as $3,000 (£1,856) from the ground up, but its solution wasn’t specifically designed for back-ups because it’s too slow, and Symantec’s software is just too cumbersome.”