Tintri's VMstore was designed to work with virtual machines (VMs) rather than physical servers. It manages and allocates storage at the VM level without using LUNs. It also performs data management functions such as snapshots, clones, replication and quality of service at the VM level.
Tintri added support for Red Hat Enterprise Virtualization last month, but customers have clamored for Hyper-V support. Support for KVM is also on the Tintri roadmap as the use of multiple hypervisors becomes more common.
"The support for Hyper-V has been a customer request. Every customer we talk to either has Hyper-V or is considering it," said Saradhi Sreegiriraju, Tintri's senior director of product management. "There certainly is a lot of demand for it. It's a pull from customers. By the end of the year we will support all three major hypervisors."
Sreegiriraju said Tintri wrote its own Server Message Block (SMB) 3.0 protocol stack that connects to Hyper-V hosts. The Tintri operating system is also integrated natively with Microsoft System Center Virtual Machine Manager and VMware vCenter.
"Our SMB stack is equal to what you get in Windows Server 2012," Sreegiriraju said.
Tintri customers can run the same appliance for VMware and Hyper-V.
According to an IDC study published in April, 59% of companies run at least two hypervisors. Hyper-V is the second most popular hypervisor behind VMware vSphere, according to the analyst firm.
"And that number [using multiple hypervisors] is increasing over time," said Eric Burgener, IDC's research director for storage.
Burgener said Tintri's approach solves the inefficiencies of trying to provision and manage storage in a server virtualization environment.
"They manage storage operations at the VM level," Burgener said. "Why is that important? All storage operations like snapshots, cloning, replication and provisioning are done on LUNs [traditional storage arrays]. That's fine if you are a storage administrator, but it doesn't help virtualization administrators."
Tintri shows a global view of all VMs that are stored and identifies performance and capacity trends without dealing with the underlying storage. It helps to identify performance hot spots at the hypervisor, network and storage level. Individual VMs are protected via customized policies for VM-level snapshots and WAN-efficient replication.