Nutanix Inc. today expanded its line of converged storage systems by launching an entry-level platform for small enterprises and branch offices, and a data center platform that handles
The new systems are the Nutanix Virtual Compute Platform NX-1000 and NX-6000. Nutanix originally used the Complete Cluster brand for its systems, but changed it to the Virtual Compute Platform to underscore its use in heavily virtual environments.
Nutanix sells storage nodes with servers, networking and hypervisors built in, eliminating the need for a separate storage network. The systems differ from converged platforms such as VCE Vblocks, EMC VSPEX and NetApp FlexPods that typically package servers, storage, networking and software, but do not integrate them into one box.
Taneja Group, a Hopkinton, Mass.-based analyst and consulting firm, uses the term "hyper-converged" for the type of systems Nutanix Inc. sells; Framingham, Mass.-based IDC recently coined the term "compustorage" to describe them. Scale Computing, Pivot3 and SimpliVity also sell systems with similar architectures.
Nutanix has another term for its type of architecture. "A lot of people think of VCE when they hear 'converged infrastructure,'" said Greg Smith, Nutanix's senior director of product and technical marketing. "But those systems are pre-engineered, pre-cabled systems assembled with legacy technology. They don't limit the complexity or the expense of legacy storage. We do the next generation of convergence. We call it 'native convergence.'"
Nutanix systems ship with VMware ESX or Red Hat KVM hypervisors built in, and the vendor plans to add the option of Microsoft Hyper-V and perhaps other hypervisors.
All Nutanix systems include Intel processors and a combination of hard drive and flash storage. Nutanix software analyzes data usage patterns, and promotes hot data to the flash and moves cold data to hard drives.
Like the NX-3000, the NX-1000 starts with four nodes in a 2U system. NX-6000 systems start with two nodes.
The NX-1050 includes 12 CPU cores, 4 TB of hard drive capacity and 400 GB of solid-state storage per node, with up to four nodes in a 2U rack. That makes for 48 cores, 16 TB of hard drives and 1.6 TB of flash in a four-node system, compared to 64 cores, 15 TB of hard drive storage and 3.2 TB of solid-state drives in the largest of Nutanix's flagship NX-3000 model.
Pricing for a four-node NX-1050 is $90,000 versus $144,000 for an NX-3000.
A 2U NX-6000 includes 32 cores, 32 TB of hard drive capacity and either 1.6 TB (NX-6050) or 3.2 TB (NX-6070) of flash. The NX-6050 costs $120,000 for two nodes, while the NX-6070 is priced at $180,000 for two nodes.
With far more storage than the NX-3000, Nutanix positions the NX-6000 systems for applications such as SQL databases, data analytics and large-scale virtual desktop infrastructure implementations that require higher capacities.
"Until now, we had one platform intended for core data centers," Smith said. "But we're getting demand for a broader set of products."
Smith said Nutanix is trying to enable a "building-block approach to building a data center" with the new systems. By adding a larger node with much greater storage capacity per node, the vendor is enabling customers to scale capacity independently of compute.
Ashish Nadkarni, IDC research director for storage systems, said the storage-intensive NX-6000 helps customers who have enough compute but need more capacity. "When you need storage, you don't necessarily always need compute," he said. "If you buy it as a packaged appliance, the compute you don't need sits there wasting cycles."
IDC created the compustorage category as a subset of converged infrastructure to distinguish systems such as Nutanix's from stacks offered by major vendors that use separate products under integrated management.
"Compustorage runs workloads natively on the storage array," Nadkarni said. "Converged systems such as Vblocks and FlexPods are delivery mechanisms put together by vendors.
"The secret sauce for companies like Nutanix is the file system and the way they allow the hypervisor to hook into the file system and run workloads natively," he said. "Other vendors are re-using their hardware box from other systems."