This Document has limited distribution rights. It may not be distributed to the general public or shared beyond the employees of the licensee company. Upon the termination of an employee, whether voluntary or involuntary, access to this document must immediately cease. If the licensee company wishes to utilize this information in outward facing communications, please contact [email protected].
On June 6, 2017, NetApp announced its first hyperconverged infrastructure (HCI) appliance, based on its all-flash SolidFire platform. This is a true HCI platform, combining multiple compute nodes, integrated networking, and a software-driven storage layer (using the SolidFire software platform). It is designed to compete against products like Dell EMC VxRAIL, Nutanix XCP, HPE SimpliVity and other similar offerings from a multitude of vendors. NetApp advised Neuralytix it is aiming the NetApp HCI at mid- to large-sized enterprises.
At its core, the software driven storage (SDS) layer is the SolidFire software platform. In practice, it is not dissimilar to Dell EMC VxRACK Flex which uses the ScaleIO software at its heart. Unlike VxRACK Flex, NetApp HCI is designed to use the VMware ecosystem to provide the hypervisor and system management.
NetApp HCI allows customers to run ONTAP Select (the software only version of ONTAP) in a virtual machine (VM), and thus providing access to the NetApp Data Fabric, which gives access to file services and object services (through the deployment of StorageGRID).
Figure 1 shows back side of NetApp HCI. The minimum configuration is two shelves. The back of each shelf provides the ability to have up to four servers inserted into the midplane. In the minimum configuration, six servers will be used – two for compute, and four for the SolidFire SDS. From a compute perspective, NetApp advised Neuralytix, that up to 100 compute nodes can be added using additional shelves. The additional shelves can be compute-only, thus overcoming the parallel scaling of compute and storage together such as that found in many other HCI solutions.
NetApp also focused heavily on simplicity, and claims that NetApp HCI is simpler to deploy than its competitors, with an end-to-end installation time of 30 minutes or less.
One of the things we found negative about the NetApp HCI approach, is that while it is heavily dependent on VMware, it does not sell the licenses. Instead, it includes a 60-day trial license. While this will give resellers and system integrators (SI) the ability to make money selling licenses, and give enterprises with enterprise licensing agreements (ELA) the ability to leverage these agreements, it also means that the support for NetApp HCI is not with a single vendor, but with at least two vendors. This is a negative in our opinion.
Our opinion of NetApp HCI is that it is clever in its ability to scale compute and storage independently. This approach leapfrogs many of the existing vendors.
During our conversations with NetApp, it continually compared itself with first generation HCI solutions. Neuralytix’s position is that all the HCI vendors have gone beyond the first generation. As such, NetApp is comparing itself to an older, out of date base standard. It is the equivalent of saying that it is a market leader in a small market – not particularly impressive.
Overall, we found NetApp HCI to be a good solution for those who are committed to the NetApp brand. We fear that NetApp HCI may not be as attractive to organizations that have an even or heavier focus on the compute as the compute nodes are as yet unproven.
Our conclusion is that the NetApp HCI is an acceptable HCI solution with the added benefit of true independent scaling between compute and storage capacities, only to have the added benefit negated because the solution and support for the solution does not come from a single source. This makes Neuralytix feel that NetApp is not as committed to converged infrastructures (as demonstrated by the reference architecture approach to FlexPod) as it is advertising.