NetApp is continuing its journey from being a legacy data storage company towards becoming a major player in enterprise AI.
At this week’s NetApp Insight customer and partner conference in Las Vegas, the firm took the wraps off its NetApp AFX hardware and software system.
The NetApp AI Data Engine that comes with the system is an extension of the well-established ONTAP data management operating system, and follows the NVIDIA AI Data Platform reference design, that helps organisations simplify and secure the entire AI data pipeline.
The scalable offering accelerates enterprise AI retrieval augmented generation (RAG) and inference across hybrid and multi-cloud environments. The system “ensures that all applicable data is immediately ready for AI”, we were told at the event. If that sounds simple, it’s supposed to.
NetApp spelt out the problems enterprises are having with their AI deployments, and it’s often mentioned across the industry that 95% of AI projects involving generative AI are failing and not being commercialised.
Companies need to clean their data pipelines to help avoid problems, said NetApp. “Brittle” data pipelines are full of data rubbish that has not been cleaned or curated properly, and that’s the challenge that NetApp now promises to solve.
“With the new NetApp AFX systems, customers now have a trusted, proven choice in on-premises enterprise storage built on a comprehensive data platform to rapidly propel AI innovation forward,” said Syam Nair, the newly installed chief product officer at NetApp. “NetApp AI Data Engine enables customers to seamlessly connect their entire data estate across hybrid multicloud environments to build a unified data foundation.
“Enterprises can then dramatically accelerate their AI data pipelines by collapsing multiple data preparation and management steps into the integrated NetApp AI Data Engine, built with NVIDIA accelerated computing and NVIDIA AI Enterprise software complete with semantic search, data vectorisation and data guardrails. The combination of NetApp AFX with AI Data Engine provides the enterprise resilience and performance built and proven over decades by NetApp ONTAP, now in a disaggregated storage architecture, and all still built on the most secure storage on the planet.”
Big words for a big problem. But key partners from NVIDIA, AWS, Microsoft, and Google were all in attendance at the show to offer NetApp encouragement in tackling the challenge of wrangling huge amounts of unstructured data, and turning it into meaningful business advantage.
NetApp AFX is an enterprise-grade disaggregated all-flash storage system, and is certified storage for NVIDIA DGX SuperPOD supercomputing to manage exabytes of data. It is designed for linear performance scaling up to 128 nodes with TBs per second of bandwidth. Optional DX50 data compute nodes enable a global metadata engine for a real-time catalogue of enterprise data, leveraging NVIDIA accelerated computing.
The NetApp AI Data Engine (AIDE) is an AI data service designed to make AI “simple, affordable, and secure”. From data ingestion and preparation to serving GenAI applications, AIDE offers a global, up-to-date view of a customer’s entire NetApp data estate for “fast” searching and curation, while “seamlessly” connecting their data to any model or tool across on-premises and the public cloud.
AIDE automates data change detection and data synchronisation, eliminating redundant copies and ensuring data is always current, helping to avoid the aforementioned brittle data pipelines.
And built-in guardrails follow data throughout its AI lifecycle, ensuring security and privacy. AIDE leverages the NVIDIA AI Data Platform reference design, featuring NVIDIA accelerated computing and NVIDIA AI Enterprise software, including NVIDIA NIM microservices for vectorisation and retrieval.
“Enterprises are looking for a trusted, high-performance data foundation to turn massive volumes of information into real intelligence that powers their AI journey,” said Justin Boitano, vice president, enterprise AI products at NVIDIA. "NetApp’s data platform has transformed into an AI-native storage platform by integrating NVIDIA accelerated computing and software, including leading AI models. With this new platform, organisations can index and search vast amounts of unstructured data across their enterprise to drive innovation and deliver real business impact.”
Long-time NetApp partner Cisco is also getting involved. The NetApp AFX architecture now integrates with Cisco Nexus switches for intra-cluster connectivity, and will “soon” extend to the partners’ combined FlexPod AI offering too.
The integration of Cisco Nexus 400G switches delivers the ultra-high bandwidth, low-latency, and “lossless” networking essential for scaling AI workloads, the partners said.
“The integration of Cisco Nexus switches into the NetApp AFX architecture exemplifies our shared commitment to AI innovation,” said Murali Gandluru, vice president, data centre networking at Cisco. “Our customers are asking for simplicity without sacrificing performance or security, which requires Cisco and strategic partners like NetApp to solve customer challenges. Together, we will move the industry forward and enable a new era of AI-driven outcomes.”
The Channel
So what about the channel in all this? Channel partners have slowly been getting their heads around the opportunity offered by AI, and a major channel player like NetApp could prove to be a useful driving force for them in the enterprise market.
Sandeep Singh, senior vice president and general manager for enterprise storage at NetApp, told IT Europa: “What we are offering can be an on-demand service as customers scale-up. And for the partners, we are offering reference architectures, use cases, and all the intel they need.
“There is excitement among our partners, and the proper channel enablement has been put in place.”
More to follow from NetApp Insight in Las Vegas...
