NetApp announces AI-Ready, Cloud-Connected, NVMe All-Flash Storage

Marcum Technology partner NetApp, recently announced AI-Ready, Cloud-Connected, NVMe All-Flash Storage. 


Self-driving cars, talking computers, and smart “tricorder” devices capable of diagnosing disease used to be the realm of science fiction. Not anymore. Fueled by vast amounts of data, recent advancements in artificial intelligence (AI) and cloud computing have enabled data visionaries to turn these fictions into reality.

Today, NetApp is announcing new products and solutions that will help data-driven organizations thrive in this new age—the industry’s fastest, most cloud-connected enterprise all-flash systems1and new data management software for hybrid clouds. They are the NVMe-accelerated NetApp®AFF A800 all-flash array and a new release of NetApp ONTAP® 9 software: ONTAP 9.4. These—along with a host of other new products and features—will enable you to deploy new AI projects, accelerate existing applications, and take advantage of industry-leading cloud services.

Designed for New AI Workloads

To meet the needs of data visionaries, NetApp is collaborating with NVIDIA, the market leader in accelerated computing and AI. Modern AI requires massive compute power for the complex neural networks that support deep learning and cognitive computing. NVIDIA’s industry-leading platform for AI uses advanced graphics processing units (GPUs) to deliver compute power that far surpasses traditional CPU-based architectures. NVIDIA GPUs require storage systems that can keep pace—systems that can manage and protect unstructured data at massive scale. The AFF A800 all-flash array sets a new bar for the industry, with an ability to linearly scale performance and feed data to NVIDIA GPU systems up to four times faster than competing solutions.2

NVMe Performance for Your Data Center

The AFF A800 supports the industry’s first end-to-end nonvolatile memory express (NVMe) architecture. It combines NVMe solid-state drives with NVMe over Fabrics (NVMe-oF) to deliver lightning-fast response times for traditional data center workloads. It also supports more I/O performance than ever before. What can you gain? How about the ability to run 60% more workloads on a single system with data access up to 50% faster than Fibre Channel Protocol?

Best of all, we’re providing this new technology in a way that protects your existing investments in SAN infrastructure. We’ve partnered with Broadcom’s Brocade storage networking division, the leading supplier of storage networking solutions, to enable Fibre Channel compatibility at the core of your data center. ONTAP 9.4 will deliver this NVMe over Fibre Channel (NVMe/FC) capability as a nondisruptive software upgrade not just on the AFF A800, but also on existing NVMe/FC-ready arrays, including the AFF A300, AFF A700, and AFF A700s.

Hybrid Cloud Data Tiering

The latest update to our flagship ONTAP 9 software also improves hybrid cloud data tiering and adds support for Microsoft Azure. ONTAP will automatically move your inactive data to a lower-cost storage tier to save money and automatically bring your data back when you need it.

In many data centers, 50% or more of the data on primary storage systems is inactive and seldom accessed, and the percentage of cold data on secondary storage is even higher. Moving that data into a lower-cost object store—whether Azure, Amazon Web Services (AWS), or internal clouds—frees up space and reduces the total amount of equipment you need to deploy and maintain in your data center. This approach not only reduces capital expenses, it saves your IT team from time-consuming and error-prone manual data management.

And that’s not science fiction.

More Information and Resources

Check out these resources to learn how these new products can help you modernize data management for the digital age.

Key AFF A800 Specifications

1 Over 1.3 million IOPS at 500 microseconds per high-availability pair; ultralow latency of less than 200 microseconds; integration with AWS, Azure, Google Cloud Platform, IBM Bluemix (IBM Cloud), and more.

2 Read throughput of up to 300GBps per all-flash cluster versus 75GBps from a leading competitor.