At the heart of the cluster is a distributed system software layer, designed specifically for virtualization, that converges compute and storage into a single tier, and is optimized to make use of flash SSDs in its core architecture. The company says that this converged architecture approach, pioneered by Google, supports virtualization capabilities like VMware's vMotion, High Availability (HA), and Dynamic Resource Scheduling (DRS), along with data management features that include VM cloning, capacity optimization, and converged backup for instant backup and recovery of virtual machine data without requiring external backup appliances.
Available now, customers can start with a single Nutanix Complete Block, a 2U unit containing four x86 nodes containing 8 Intel Xeon processors and 192 GB RAM (upgradable to 768 GB) along with 1.3 TB of Fusion-io, 1.2 TB of SATA SSD, and 2 0TB of SATA drives, and running an industry-standard hypervisor (VMware ESXi). Deployment takes less than 30 minutes, and expansion just requires adding another block, which start at $75,000 each.
The biggest strength of the cluster is the elimination of the network bottleneck between the data and the CPUs, said Aberdeen analyst Dick Csaplar. "With the data stored internally in the server, the computing process can be kept refreshed and up to date more efficiently than having to make calls to an independent and remote storage device."
He said dealing with storage and computing as a single tier is being done, but not in a commercially offered, prepackaged solution targeting the mid to large enterprise. "I don't know of any other start-ups focused on this area, as most companies take an on-site SAN or NAS as a given." Csaplar added that Nutanix can find real traction for their targeted use cases, as eliminating the need for expensive SAN devices will always interest people.
InformationWeek Analytics has published a report on backing up VM disk files and building a resilient infrastructure that can tolerate hardware and software failures. After all, what's the point of constructing a virtualized infrastructure without a plan to keep systems up and running in case of a glitch--or outright disaster? Download the report now. (Free registration required.)