- 18 Hot IT Certifications for 2014
- CIOs Opting for IT Contractors Over Hiring Full-Time Staff
- 12 Best Free iOS 7 Holiday Shopping Apps
- For CMOs Big Data Can Lead to Big Profits
Network World - When Tata Communications, the Indian telecommunications company, rolls out its infrastructure as a service cloud offering in America in the coming months, company officials want to claim differentiating features in the products. And one they're hoping to include is support for multiple hypervisors.
While most cloud infrastructure service providers architect the data centers that host their cloud offerings around a single hypervisor format, such as VMware or Citrix's XenServer, a growing trend in the cloud industry, according to some analysts, is around service providers offering customers their choice of the underlying virtualization software powering the cloud offering.
"Customers are multi-hypervisor, so we strive to support them," says Tata Senior Vice President John Landau.
But from a customer perspective, does it really matter what hypervisor a service provider uses?
"From a technology standpoint, it shouldn't really matter," says Agatha Poon, who tracks the global cloud market at the 451 Research Group. She says along with providers moving up the stack into providing platform-as-a-service offerings, she believes providers will look to offer multiple hypervisors as well.
Hypervisors have been fairly commoditized and various brands, including those from VMware, Citrix, as well as other formats such as KVM and Microsoft's Hyper-V, offer similar feature sets, she says. But there could be some reason for customers to care which hypervisors their provider uses.
Specifically, if a customer is looking to migrate workloads between their on-site environment and the public cloud, then having common hypervisor software could be helpful, says cloud analyst Philbert Shih of Structure Research. It can reduce the need to rewrite applications to operate with specific hypervisors, for example.
But, if a company uses the cloud for new applications that are deployed in the cloud and will live there for their entire life cycle, then it shouldn't really matter the hypervisor of the service provider, he says.
Having portability of workloads between the private and public clouds is exactly what Allan Leinwand was looking for when he architected the hybrid cloud that powers Zynga, the social gaming site. Leinwand's approach is to "own the base and rent the spike," he said during a keynote speech at Interop this week.
That means that Zynga's infrastructure, dubbed zCloud, has enough on-site capacity to handle most of the company's common IT needs. When there is a spike in demand, from a new game being launched, to one going viral, zCloud uses public cloud resources. In order to make for easy migration of workflows between the on-premise zCloud and the Amazon Web Service (AWS) public cloud, Leinwand constructed the zCloud to look, feel and act as much like the AWS cloud as possible, he says. Part of that involved using the same hypervisor equipment as AWS, which is the Citrix Xen hypervisor.