NetApp has introduced its Novus storage architecture, designed to support AI data infrastructure by providing a unified approach to storage and management. The system achieves an aggregate throughput exceeding 100TB/s, with the Data Director component managing metadata independently from stored data to allow for scalable operations.
The architecture provides applications with a unified view of files across multiple ONTAP storage clusters. Additionally, the company's AI Data Engine is designed to automate the discovery, classification, and vectorization of enterprise data, aiming to reduce data preparation timelines that typically span six to nine months.
To integrate with existing enterprise environments, NetApp is utilizing open standards and plans to acquire PEAK:AIO Ltd. This acquisition will incorporate parallel file system technology based on the pNFS protocol, which is supported by Linux kernels released after 2018. The company emphasizes that this infrastructure is intended to work alongside current systems without requiring a complete replacement.
