Hewlett Packard Enterprise (HPE) is overhauling its Ezmeral software program portfolio and simplifying knowledge group, administration and analytics for organizations advancing their synthetic intelligence (AI) and machine studying (ML) workflows.

HPE Ezmeral, HPE’s enterprise software program division, has grown lately by way of a sequence of acquisitions. In 2018, HPE acquired BlueData software, which introduced in new capabilities for utilizing utility containers to handle knowledge in help of AI/ML efforts. In 2019, it acquired knowledge platform vendor MapR, which introduced capabilities for the Hadoop Distributed File System (HDFS). In 2021, HPE acquired Ampool, including a distributed SQL engine for knowledge queries. And on Could 2 HPE employed a lot of the management and engineering workforce behind the recently-acquired Arrikto, one of many main contributors to the open-source Kubeflow ML workflow project. HPE Ezmeral has additionally been supporting a number of different open-source efforts which are essential to trendy knowledge and AI operations, together with Apache Spark, Apache Airflow, Apache Superset, MLflow, Feast and Ray.

HPE Ezmeral introduced right this moment that it’s now organizing and updating its portfolio with two main streams. The brand new HPE Ezmeral Data Fabric software program suite will consolidate the corporate’s knowledge administration applied sciences. HPE Ezmeral Unified Analytics Software program is the brand new offering that will bring together all the tools and technologies needed for organizations to manage data analytics in addition to AI/ML workloads.

“We’re introducing the brand new simplified portfolio the place we’ve chosen to principally go away from doing issues like constructing out bespoke options for container platforms, MLOps and Spark,” Mohan Rajagopalan, vice chairman and normal supervisor at HPE Ezmeral Software program, advised VentureBeat. “We need to concentrate on offering basis capabilities that assist our prospects develop and deploy purposes in a hybrid multicloud world.”

Why HPE is simplifying the Ezmeral software program portfolio

Rajagopalan defined that earlier firm acquisitions had meant persevering with with the acquired firms’ expertise, beneath the HPE umbrella.

“Right now, what we’re doing is we’re principally taking a step again and how these varied applied sciences play collectively and extra importantly, spending time with our prospects to know the place we needs to be enjoying,” Rajagopalan mentioned.

What is clear to HPE and lots of others now could be that knowledge is essential to success for contemporary enterprises.  Knowledge helps to drive decision-making with analytics and enterprise intelligence. Knowledge can also be the inspiration for AI/ML.

“Virtually each buyer says that knowledge is the long run — and whoever owns extra knowledge has higher insights,” he mentioned. “The expertise to generate perception goes to maintain evolving over time.”

Having knowledge is about having the precise instruments to handle knowledge and make sense of it. It’s additionally about having the infrastructure that helps knowledge initiatives. To that finish, Rajagopalan famous that HPE Ezmeral enhances the HPE Greenlake portfolio. Ezmeral is solely centered on the software program layers of the stack, whereas Greenlake takes extra of an infrastructure view.

Why knowledge cloth is vital

Knowledge in a contemporary enterprise is usually present in many alternative places and purposes. With the HPE Ezmeral Knowledge Cloth Software program, the aim is to supply a collection of capabilities to assist enterprises manage, handle and join knowledge.

Rajagopalan defined that with Knowledge Cloth, HPE is offering capabilities for knowledge present in information, streaming knowledge sources and databases. The essential concept is to allow a federated knowledge layer, the place completely different sources might be related in an method that permits organizations to profit from all their knowledge.

Functionally, what Knowledge Cloth is enabling is a hybrid data lakehouse, the place knowledge might be saved on-premises, on the edge or within the cloud and is obtainable for knowledge evaluation.

“We create knowledge lakehouses the place prospects can merely pump knowledge in a wide range of codecs into the material. It may be several types of knowledge and it may be in numerous places,” Rajagopalan mentioned. “Knowledge Cloth simply makes the magic occur the place the information seems the place it must be consumed.”

Knowledge by itself isn’t sufficient to be helpful. That’s the place the HPE Ezmeral Unified Analytics Software program suite comes into play.

“What we’ve accomplished is we’ve principally taken the best-of-breed open-source applied sciences, so take into consideration Superset, Kubeflow, Airflow, Feast and Ray, and we bundle them with enterprise-grade guardrails beneath this umbrella referred to as Unified Analytics,” Rajagopalan mentioned.

With all the thrill and hype round giant language fashions (LLMs) and the runaway success of OpenAI’s ChatGPT, many organizations want to profit from AI. Whereas the most important LLMs require large quantities of knowledge, Rajagopalan emphasised that there’s appreciable worth within the knowledge that enterprises have already got that would probably be used for constructing out AI fashions.

“We need to present all of them the instruments to allow them to begin constructing out their very own particular domain-specific fashions. They could possibly be LLMs, they could possibly be chat-focused, they could possibly be workflow-focused,”  Rajagopalan mentioned. “Our thesis right here is that enterprises are sitting on treasure troves of knowledge.”