HPE updates Alletras X and B10000, Zerto and Data Fabric in GreenLake private cloud update blast

HPE has updated its Alletra X10000 with a native file system and announced AI assistance for the Alletra B10000 object system, Zerto continuous data protection and Data Fabric data moving software.

HPE”s private cloud GreenLake file storage was originally announced back in 2023 with the Alletra 10000 disaggregated storage array powered by OEM’d VAST Data file-based storage software. Now HPE has added its own ground-up-developed filesystem software. This is, as SVP and GM of HPE Storage Jim O’Dorisio, said: “alongside the existing object storage capability in the same platform.” Native file support “gives us the ability to design and architect the kind of tight integration to deliver on what we really envision is the data services layer for the hybrid cloud and to deliver on the kinds of outcomes that our customers are looking for from an intelligence capability.”

The filesystem supports file RDMA (remote direct memory access) which adds to the prior support for S3 (object storage) over RDMA. Both the file and object stores share the same foundation; O'Dorisio again: “the entire unstructured platform is built on a key value store. So we architected that from the ground up to be an intelligence platform. We started off with object, we've delivered file this month. … [They are] equal citizens with their own namespaces on the unstructured platform [and] the initial release of the file capability literally has performance on par with our object store.”

The filesystem is POSIX-compliant and supports NFS v4.1. But, O'Dorisio explained: “We're not going after, at least not initially, the traditional filer environment per se. We're really targeting … the intelligence workloads and our partnerships with Nvidia really reflects that.” There is no need for HPE’s VAST-based filesystem customers to update their systems as both options will be supported.

O’Dorisio believes that: “hybrid [file +object] is really becoming the predominant deployment mode in enterprises. And quite frankly, we see AI workloads as being the predominant hybrid workload.”

The native HPE software for the X10000 “is literally a Kubernetes cluster. And so all of our compute nodes and storage nodes are architected into a disaggregated, scale-out architecture, but it's a Kubernetes cluster. And we recently released the Data Intelligence Node. … That Data Intelligence Node allows us to actually calculate vector embeddings as the data lands on the platform and we house our own vector database within the X10000.”

The X10000, which scales to 16 compute and 16 storage nodes, and 23PB raw capacity, now has a 100 percent data availability guarantee to keep applications continuously running without disruption. It also supports MCP.

AI Assistants

Its Alletra Storage MP twin system, the B10000 has been given real-time, agentic support with a v10.6.0 SW release, to autonomously detect, analyze, and resolves infrastructure issues. This moves “beyond signature-based predictive analytics and pattern matching with semantic understanding, adaptive reasoning, and prescriptive intelligence.” It’s designed to anticipate and resolve issues autonomously earlier in the support lifecycle, instead of reacting after issues escalate.

This feature relies on a coordinated set of specialized AI agents:

• A detection agent that identifies anomalies in real time
• An analysis agent that performs root‑cause investigation
• A recommendation agent that prescribes targeted corrective actions
• A remediation agent capable of implementing autonomously by creating approved fixes

These agents draw insight from system telemetry metadata, best practices, and accumulated product knowledge across the installed base.

There is also “a 5:1 data reduction guarantee to increase efficiencies and lower overall cost per TB.” And the B10000 can now scale from four to six controller nodes, with single node increments, to provide 50 percent more performance and has built-in dual-node fault tolerance. 

AI assistance has been added to HPE’s Zerto continuous data protection software, and Zerto has been integrated with Microsoft Defender to deliver real-time threat visibility and faster recovery. 

HPE has an X10000-based Data Fabric capability that supports distributed edge sites, co-lo centers, core data centers and the three main public clouds. It has a global namespace, universal access, multi-protocol support, automated tiering, caching, and mirroring. V8.1 of the Data Fabric software features;

• Policy-based data placement and movement (tiering), 
• Conversational interface and agentic AI assistant for faster insights, automated reporting, and improved decision-making. 
• Enhanced metadata integration to improve visibility, classification, and lineage, 
• Support for the Apache Polaris catalog for Iceberg tables, for consistent governance and compliance across platforms. 

It supports real-time movement of objects between any S3-compatible storage systems meaning that data can be moved or copied to external S3 buckets. AI teams can ingest new data from S3-backed sources into a governed Data Fabric Software environment faster, without relying on slow, manual batch transfers. It makes AI inferencing processes faster.

Explore the X10000 filesystem news here. Read more about the B10000 news here. Learn more about the Data Fabric SW release here.