HPE Solutions with WEKA

Unlock Maximum Speed with AI-Native Storage

Increase utilization of your valuable GPUs by avoiding the storage bottleneck. HPE Solutions with WEKA delivers scalable, AI-native storage with high throughput, IOPS, and ultra-low latency that meets the rigorous standards of NVIDIA Cloud Partner (NCP) certification.

Slash AI job runtimes with the world’s fastest shared file system (SPEC Storage 2020, SPEC SFS 2014, IO-500, and STAC). Process billions of files at unbeatable speeds, whether handling small files or mixed workloads. Reach peak efficiency and I/O with full NVIDIA Magnum IO™ software suite support.

Make data copying and storage silos a thing of the past. Run any part of your pipeline on a single system, regardless of data size, type, or volume. The unique zero-copy, parallel architecture distributes metadata across nodes, making data as fast as local across all types and patterns.

Unleash WEKA’s ultra-fast file performance on HPE’s world record-holding¹ PCIe Gen5 all-NVMe servers. Optimized for AI/ML scenarios, these servers are powered by the latest processors from AMD and Intel®, including NCP-qualified configurations.

Don’t let the network slow down your AI training and inferencing. Get high-speed connectivity of up to 400 Gbps Ethernet or InfiniBand network cards and HPE’s fast network switch offerings from NVIDIA and others.

Gain Transformative Scale Wherever Your Data Pipeline Lives

Scale your AI storage to hundreds of petabytes—or even exabytes—across tiers, on-prem, cloud, or hybrid environments. This future-driven, multi-tenant, multi-workload, multi-location platform effortlessly handles billions of files in a single namespace.

Mitigate management complexity with a unified platform where all data sets live in a single namespace. Automatically managed data tiering streamlines file management, sharing, and access, while GPUs gain rapid access to data needed for training.

Easily scale to 512 PB of flash. Extend file system namespace to scale to exabytes with an economical object tier through HPE Solutions with Scality or any S3-compatible object store, public, or private.

Seamless multi-tenancy enables easy management of multiple customers, teams, or projects at desired privacy, security, and performance levels with composable clusters on shared hardware. Scale as priorities evolve without impacting the entire cluster, optimizing costs and management overhead.

Deploy flexibly across multiple environments (bare-metal, containers, native in cloud) with built-in multi-protocol access at no extra charge. Simultaneous NVIDIA GPUDirect Storage (GDS), POSIX, S3, NFS, SMB, and HDFS support enable a single platform to handle all your AI data lifecycle needs.

Simplify Operations with Fast File Storage Built for Enterprise

Take the complexity out of fast file storage. Management is made easy, requiring no tuning or storage specialists. Modernize with AI/ML-optimized hardware offering superior quality and security. With a full best-in-class solution, intensive administration is history—just set it and forget it.

Auto-adapt to any workload instantly. Built-in intelligent data management helps ensure peak performance without manual tuning. Effortlessly configure, monitor, and administer clusters from one place with WEKA Home’s full-featured, intuitive console—no IT expert needed.

Enable fast, secure data access for multiuser enterprises. Protect against rogue actors with end-to-end encryption (at-rest and in-transit), authentication, and automatic cloud backups. The unique snap-to-object feature for disaster recovery safeguards your data without performance loss.

Protect your AI workloads with best-in-class servers and availability. High-load servers with built-in redundancy and failovers reduces downtime, while silicon root of trust anchors firmware. Advanced monitoring and behavioral analytics deliver robust end-to-end IT protection.