“Data Driven” was a key theme at NetApp Insight in Las Vegas last week. According to CEO George Kurian and other executives that spoke, becoming data driven requires legacy IT to transform, using the cloud – more specifically, a hybrid, multi-cloud architecture. This means building a data strategy that moves from batch to real time processing, from data centers to data fabrics and from fragmented to centralized. More on these concepts below. Kurian also emphasized that “speed is the new scale”, referring to the importance of agile infrastructure and agile thinking. That said, companies must still be able to deal with increasingly larger data sets, like Dreamworks, a big NetApp customer.
Dreamworks, the film and digital media company, shared some of the realities of being a data-driven company – one that creates an enormous amount of data. Feature length movies, like “How to Train your Dragon”, (for which they showed a pre-release clip), have about 130K frames. This adds up to about ½ billion active files that must be available, protected and managed throughout their 12-step production process. Dreamworks has about 10 films in process at any one time, meaning their infrastructure must support 5 billion active files.
NetApp’s Data Fabric lets them move these between on-prem infrastructure and multiple cloud-based systems. Data Fabric is NetApp’s data architecture that enables data to be managed and seamlessly transferred between on-prem and cloud-based infrastructures. It provides a consistent set of data services and tools to ensure access, control, protection and security for data on-prem or in the cloud.
The most prominent infrastructure solution at Insight was NetApp’s HCI. One major difference between this and other HCIs is the NetApp HCI disaggregated architecture, which runs the storage and compute functions on separate nodes. Most HCIs combine these functions on each node using a software defined storage layer that runs as a VM. The disaggregated architecture allows NetApp HCI to scale storage and compute independently and takes storage processing tasks off the CPUs that run compute functions. For more information, see the Product Brief for NetApp HCI.
NetApp presented the HCI product as the ideal infrastructure for the on-prem portion of a hybrid cloud – with connections to one or more public clouds. Underscoring this emphasis NetApp introduced this product as the “Hybrid Cloud Infrastructure” (not “hyperconverged infrastructure”). This may have also been in response to some in the industry saying disaggregated architectures aren’t HCIs, a position not shared by Evaluator Group and detailed in a recent blog.
Cloud Volumes Service are storage volumes that can be set up from a single pane of glass and run in the public cloud (AWS or Google Cloud). Managed and sold by NetApp on a capacity basis, they have the look and feel of on-prem NetApp storage that’s resident in the cloud, not just buckets of capacity. With Cloud Volumes ONTAP, companies can also set up their own Cloud Volumes running ONTAP software in the public cloud and pay for the resources consumed directly to the cloud provider.
Kubernetes is becoming the de facto container orchestration platform. With the acquisition of StackPointCloud, a Kubernetes as a Service offering, NetApp last month released NKS, providing “3-click, cloud-based provisioning” to deploy and manage Kubernetes clusters across multiple clouds and on-prem NetApp infrastructure (NetApp HCI). NKS uses the NetApp Trident Plugin for Docker and Kubernetes to provide persistent storage management functions.
Cloud Insights is a SaaS-based infrastructure monitoring tool that uses advanced analytics enabling companies to manage resources across public clouds and on-prem to maximize uptime and optimize their resources. Cloud Central is the single pane of glass that brings cloud storage (Cloud Volumes), data services (Cloud Sync, Tiering, Backup, etc), NKS and platform solutions (MongoDB) together.
ONTAP AI is a rack-based, converged infrastructure solution that’s purpose built for AI and Deep Learning applications. It includes NetApp All-flash FAS storage with Cisco switches and NVIDIA DGX supercomputer servers connected via RDMA over RoCE. ONTAP AI was released shortly before Insight. From “Memory Accelerated Data”, MAX Data is a persistent memory cache software solution that came from NetApp’s Plexistor acquisition in 2017. It runs on commodity server hardware and uses 3D XPoint Optane devices to provide single-digit microsecond latencies, auto-tiering data to an ONTAP array using NVMe over FC. MAX Data is part of the ONTAP 9.5 release.
At Insight, the emphasis was on cloud, data transformation and using data as a competitive advantage, themes common at all other infrastructure vendors’ events. However, NetApp seems to be ahead of most other vendors in delivering on this vision with Data Fabric tying on- and off-premises infrastructures together in a meaningful way.
NetApp needs to get the market to understand the value of their HCI solution, especially in the larger, higher-performance use cases that enterprises have. This is an area that the rest of the HCI vendors have been trying to penetrate, without huge success. If NetApp can get these organizations to include their product in HCI evaluations we could see a different result.
NetApp’s choice to re-define the HCI acronym to mean “Hybrid Cloud” Infrastructure (instead of hyperconverged) may have been a shot at some in the analyst community that refuse to include this product in their comparisons. The proof of this pudding will be in the value of disaggregated architectures (this is clearly a direction that HCI is moving) and in NetApp’s ability to leverage the Data Fabric and Cloud Volumes technologies to make HCI the on-prem portion of the enterprise hybrid cloud.
NetApp had a lot of very happy customers and partners in attendance at Insight 2018. The company continues to show strong performance in the market with 30% y/y growth and a position as the leading provider of all-flash (AF) storage. They are also #1 in AF in Japan. These numbers reflect those customers’ continued preference for NetApp’s existing technology to a large extent. As they move further down the road with AI, digital transformation and multi-hybrid clouds, the company needs to needs to make sure they don’t leave their on-premises customers behind.