Enterprises want an entire lot of knowledge to energy their AI purposes. That features structured and unstructured knowledge, block and file knowledge, streaming and static knowledge, and knowledge within the cloud and on the sting. To assist preserve all that knowledge straight, Hewlett Packard Enterprises at the moment rolled out a reimagining of its HPE Information Cloth providing. It additionally bolstered its Alletra storage home equipment and launched a brand new GPU compute providing dubbed Nvidia AI Computing by HPE.
HPE debuted its first knowledge lakehouse answer again in 2021 when it launched the Ezmeral Information Cloth Object Retailer as a part of its GreenLake Cloud Service. The objective of the software program was to offer a standardized method to ingest, retailer, and serve the massive number of knowledge sorts that enterprises needed to make use of for analytics.
Whereas analytics remains to be vital, what’s actually driving the needle in huge knowledge at the moment is AI in addition to agentic AI. To that finish, HPE at the moment made a collection of bulletins round its “new unified knowledge layer,” with the said objective of serving to clients ship on their AI methods. Particularly, HPE introduced that its knowledge cloth now works with Alletra Storage MP X10000 equipment.
HPE has had the information cloth product available in the market for quite a few years and has quite a lot of joyful knowledge lakehouse clients, mentioned Jim O’Dorisio, the SVP and GM of HPE Storage, in a press briefing final week.
“It’s been very efficient at that,” he continued. “However more and more, we’ve constructed out the worldwide namespace functionality of the product that helps us cope with this distributed knowledge problem from edge to cloud. And it’s not solely distributed, it’s heterogeneous. So this capability to offer a single namespace from edge to cloud throughout a heterogeneous set of knowledge sources to ship common entry multi-protocol assist, automated tiering and safety is totally foundational to enabling knowledge.”
As a part of the mixing of the HPE Information Cloth Software program with its MP X10000 equipment, HPE is delivering a brand new, automated, inline metadata tagging engine that it says is designed to create AI-ready object knowledge.
“It is a actually cool functionality,” O’Dorisio mentioned. “It permits clients to actually chat with their knowledge virtually instantly upon ingestion. And that is very distinctive as a result of we’re truly enriching the information as we’re ingesting the information.”
HPE additionally supplied an replace on the plan it introduced in November to work with Nvidia so as to add assist for Distant Direct Reminiscence Entry (RDMA) between X10000 and Nvidia.
“We’ve accomplished that performance,” O’Dorisio mentioned. “Will probably be rolled out on this coming launch, and this gives as much as 6x efficiency enchancment from each a bandwidth and a latency perspective by way of RDMA interfaces.”
HPE launched the Alletra Storage MP collection, you’ll bear in mind, two years in the past to offer expandable storage home equipment as a part of the GreenLake cloud service. HPE tapped its companion VAST Information to offer the file storage functionality, to associate with the Alletra equipment’s present assist for bock storage.
HPE Alletra MP
HPE additionally made a number of enhancements to Alletra Storage MP B10000, the block storage arrays that it sells as a part of the GreenLake Cloud. For starters, HPE added file entry to the MP B10000 arrays to associate with the block functionality. That led HPE to assert that it’s the primary and solely vendor “to assist disaggregated scale-out unified block and file on a single working system (OS) and storage structure.”
HPE additionally delivered a brand new service dubbed HPE Alletra Block Storage for Azure, which provides block storage capabilities for hybrid cloud environments on Microsoft Azure. It additionally added assist for ransomware safety by way of companion Zerto.
Individually, HPE introduced enhancements to HPE Non-public Cloud AI. For starters, it’s bolstering its GPU compute providing, dubbed Nvidia AI Computing by HPE, which the corporate launched final yr and which it says encompasses “a full-stack, turnkey personal cloud for AI.” This atmosphere will get the newest Nvidia GPUs, together with Blackwell and Grace Blackwell chips, and Nvidia AI software program, the corporate says.
“We’re maniacally targeted on time to market with the newest Nvidia innovation,” Fidelma Russo, HPE CTO, government vice chairman, and GM of Hybrid Cloud, mentioned throughout a press briefing final week. “The true perform of that is to be sure that we are able to ship the expertise that’s fueled by AI to our clients…and present them get financial worth out of these applied sciences.”
HPE additionally bolstered its HPE Non-public Cloud AI providing with a number of different updates, together with assist for the aforementioned HPE Information Cloth. It additionally good points a brand new developer system that features “an built-in management node, end-to-end AI software program and 32TB of built-in storage.”
Supporting AI workloads on the sting is a should for HPE, particularly relating to AI inferencing workloads, CTO Russo mentioned on the press name. “As everyone knows, we’re within the age of GenAI,” Russo mentioned. “We’ve began with coaching, and we’ve seen…great acceleration of coaching throughout the board during the last couple of years. We’re now beginning to strategy the place the place inferencing goes to grow to be an even bigger a part of the story, and particularly inferencing on the edge.”
Lastly, HPE Non-public Cloud AI additionally will get assist for Nvidia “blueprints,” which provides clients entry to a library of pre-built AI software program, particularly the Multimodal PDF Information Extraction Blueprint and Digital Twins blueprints. HPE introduced it’s bettering its observability capabilities in HPE OpsRamp by including assist for GPU optimization by way of HPE Non-public Cloud AI.
HPE made these bulletins as Nvidia’s GPU Expertise Convention (GTC) 2025 will get underway in San Jose, California.
Associated Objects:
HPE Faucets VAST Information For Quick File Storage on GreenLake
HPE Provides Lakehouse to GreenLake, Targets Databricks
HPE Cranks the Density on GreenLake NAS
The put up HPE Preps for the AI Period with Up to date Information Cloth, Storage, and Compute Choices appeared first on BigDATAwire.
