NetApp sets direction firmly towards AI & ML workloads with new strategy

CEO George Kurian has detailed NetApp's strategic direction in advancing AI innovation through sophisticated data infrastructure today. Kurian highlighted the importance of enabling a unified and integrated generative AI (GenAI) stack, emphasising that success in AI hinges on managing governable, trusted, and traceable data.

NetApp has announced progress in several key initiatives designed to enhance AI capabilities. These include the NVIDIA DGX SuperPOD Storage Certification for NetApp ONTAP. This certification, running on the AFF A90 platform, will enhance data management for large AI projects. Kurian noted that this development would eliminate the need for compromises in data management for AI training workloads.

The company is also exploring the creation of a global metadata namespace, which aims to manage data securely and compliantly across hybrid multi-cloud estates. This will enable effective feature extraction and data classification for AI. A new integration with NVIDIA AI software, leveraging this global metadata namespace, will reportedly enable enhanced enterprise retrieval augmented generation (RAG) for AI.

NetApp's innovations further include a directly integrated AI data pipeline. This pipeline aims to prepare unstructured data for AI by iteratively capturing incremental changes, classifying and anonymising data based on policy, and creating highly compressible vector embeddings. The embeddings will be stored in a vector database integrated with the ONTAP data model. This setup is designed to facilitate high-scale, low-latency semantic searches and RAG inferencing.

Additionally, NetApp is introducing a disaggregated storage architecture that allows for full sharing of the storage backend. This architecture aims to optimise the utilisation of network and flash speeds while reducing infrastructure costs. It is intended to improve performance for high-scale, compute-intensive AI workloads, such as large language model (LLM) training, while maintaining ONTAP's data management, security, and governance features.

The company is expanding its native cloud services to drive AI innovation in the cloud. NetApp's cloud services will provide a centralised data platform for ingesting, discovering, and cataloguing data. Integration with data warehouses and developing data processing services will facilitate the visualisation, preparation, and transformation of data for AI and machine learning applications. Planned integrations with Google Cloud will enable customers to use Google Cloud NetApp Volumes for BigQuery and Vertex AI.

"Organisations of all sizes are experimenting with GenAI to increase efficiency and accelerate innovation," said Krish Vitaldevara, Senior Vice President at NetApp. He added, "NetApp empowers organisations to harness the full potential of GenAI to drive innovation and create value across diverse industry applications. By providing secure, scalable, and high-performance intelligent data infrastructure that integrates with other industry-leading platforms, NetApp helps customers overcome barriers to implementing GenAI."

NetApp continues strengthening partnerships within the AI ecosystem. Domino Data Labs has chosen Amazon FSx for NetApp ONTAP to enhance machine learning operations (MLOps). This partnership underscores the importance of seamless integration in AI workflows. Direct integration between Domino's MLOps platform and NetApp ONTAP is also underway to streamline data management for AI workloads.

General availability has been announced for AIPod with Lenovo for NVIDIA OVX, a converged infrastructure solution aimed at enterprises leveraging generative AI and RAG capabilities. New features have also been added to NetApp's FlexPod AI solution, which seeks to simplify, automate, and secure AI applications using a hybrid infrastructure and operation platform, incorporating Cisco compute and network technologies alongside NetApp storage.

"Implementing AI requires finely tuned pieces of technology infrastructure to work together perfectly," said Mike Leone, Practice Director at Enterprise Strategy Group. "NetApp delivers robust storage and data management capabilities to help customers run and support their AI data pipelines."