Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More
Hewlett Packard Enterprise (HPE) is overhauling its Ezmeral software portfolio and simplifying data organization, management and analytics for organizations advancing their artificial intelligence (AI) and machine learning (ML) workflows.
HPE Ezmeral, HPE’s enterprise software division, has grown in recent years through a series of acquisitions. In 2018, HPE acquired BlueData software, which brought in new capabilities for using application containers to manage data in support of AI/ML efforts. In 2019, it acquired data platform vendor MapR, which brought capabilities for the Hadoop Distributed File System (HDFS). In 2021, HPE acquired Ampool, adding a distributed SQL engine for data queries. And on May 2 HPE hired most of the leadership and engineering team behind the recently-acquired Arrikto, one of the leading contributors to the open-source Kubeflow ML workflow project. HPE Ezmeral has also been supporting multiple other open-source efforts that are critical to modern data and AI operations, including Apache Spark, Apache Airflow, Apache Superset, MLflow, Feast and Ray.
HPE Ezmeral announced today that it is now organizing and updating its portfolio with two primary streams. The new HPE Ezmeral Data Fabric software suite will consolidate the company’s data management technologies. HPE Ezmeral Unified Analytics Software is the new offering that will bring together all the tools and technologies needed for organizations to manage data analytics as well as AI/ML workloads.
“We are introducing the new simplified portfolio where we’ve chosen to basically go away from doing things like building out bespoke solutions for container platforms, MLOps and Spark,” Mohan Rajagopalan, vice president and general manager at HPE Ezmeral Software, told VentureBeat. “We want to focus on providing foundation capabilities that help our customers develop and deploy applications in a hybrid multicloud world.”
Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.
Why HPE is simplifying the Ezmeral software portfolio
Rajagopalan explained that previous company acquisitions had meant continuing with the acquired companies’ technology, under the HPE umbrella.
“Today, what we’re doing is we’re basically taking a step back and looking at how these various technologies play together and more importantly, spending time with our customers to understand where we should be playing,” Rajagopalan said.
What is obvious to HPE and many others now is that data is critical to success for modern enterprises. Data helps to drive decision-making with analytics and business intelligence. Data is also the foundation for AI/ML.
“Almost every customer says that data is the future — and whoever owns more data has better insights,” he said. “The technology to generate insight is going to keep evolving over time.”
Having data is about having the right tools to manage data and make sense of it. It’s also about having the infrastructure that supports data initiatives. To that end, Rajagopalan noted that HPE Ezmeral complements the HPE Greenlake portfolio. Ezmeral is purely focused on the software layers of the stack, whereas Greenlake takes more of an infrastructure view.
Why data fabric is important
Data in a modern enterprise is often found in many different locations and applications. With the HPE Ezmeral Data Fabric Software, the goal is to provide a suite of capabilities to help enterprises organize, manage and connect data.
Rajagopalan explained that with Data Fabric, HPE is providing capabilities for data found in files, streaming data sources and databases. The basic idea is to enable a federated data layer, where different sources can be connected in an approach that allows organizations to benefit from all their data.
Functionally, what Data Fabric is enabling is a hybrid data lakehouse, where data can be stored on-premises, at the edge or in the cloud and is available for data analysis.
“We create data lakehouses where customers can simply pump data in a variety of formats into the fabric. It can be different types of data and it can be in different locations,” Rajagopalan said. “Data Fabric just makes the magic happen where the data appears where it needs to be consumed.”
Unified analytics is powered by open-source tools
Data on its own isn’t enough to be useful. That’s where the HPE Ezmeral Unified Analytics Software suite comes into play.
“What we’ve done is we’ve basically taken the best-of-breed open-source technologies, so think about Superset, Kubeflow, Airflow, Feast and Ray, and we package them with enterprise-grade guardrails under this umbrella called Unified Analytics,” Rajagopalan said.
With all the excitement and hype around large language models (LLMs) and the runaway success of OpenAI’s ChatGPT, many organizations are looking to benefit from AI. While the biggest LLMs require massive amounts of data, Rajagopalan emphasized that there is considerable value in the data that enterprises already have that could potentially be used for building out AI models.
“We want to provide them all the tools so they can start building out their own specific domain-specific models. They could be LLMs, they could be chat-focused, they could be workflow-focused,” Rajagopalan said. “Our thesis here is that enterprises are sitting on treasure troves of data.”
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.