A ‘Herculean’ HPE Ezmeral Software AI Makeover: 7 Things You Need To Know

HPE released a new version of its Ezmeral Software data fabric that transforms the SaaS offering into a much more powerful platform for AI artificial intelligence and ML based workloads.

New HPE Ezmeral Software Tuned For AI And ML Workloads

Hewlett Packard Enterprise has completed a ‘Herculean’ makeover of its Ezmeral Software data fabric that transforms it into a much more powerful platform for AI and ML based workloads.

The HPE makeover was made to make HPE “more relevant” with AI and ML workloads with a focus on making it easier and more cost effective to install, build and deploy AI and ML applications, said HPE Ezmeral Worldwide Software Vice President of Sales Steve Shattuck.

Shattuck credited HPE Vice President and General Manager of Ezmeral Software Mohan Rajagopalan and his 800 to 1000 member strong software development team for making dramatic improvements that will make the edge to cloud data fabric much more attractive to customers looking to deploy AI and ML workloads.

“The number one thing the engineering development team heard from the field is ease of use, we have got to make it easier for our customers to deploy and consume,” said Shattuck. “Mohan made it the north star to make this dramatically and radically easier to use. Nobody in the marketplace is going to have a product with this much breadth that is as easy to install. It was a Herculean effort.”

The makeover essentially takes what were four HPE Ezmeral product lines, collapsing them into two products:the HPE Ezmeral Software data fabric and a new Unified Analytics product.

The Unified Analytics product combines the old unified analytics and machine learning operations product into a single product that spans the complete unified analytics/machine learning experience from creating AI models to production AI applications.

The new offering has simplified what was a number of different offerings built from the HPE acquisitions of BlueData in 2018, MapR assets in 2019, Scytale in 2020 and Ampool in 2021.

In the past, HPE would have to get partners to go through “weeks of training” to be able to install and administer the HPE Ezmeral Software, said Shattuck. “That weeks of training is now going down to hours to days in terms of their ability to deploy the software,” he said.

HPE has effectively “tripled” the ability to on board partners, said Shattuck. “The lifting that it is going to take to get partners ready to sell is going to be so much less,” he said. “Before we would have to get weeks of commitment from a channel partner to go through all the training to install the software, then go through more training on how to administer the software before they were ready to actually engage in a project. Now we can onboard a partner in implementation in a day and then the selling proposition in another day.”

HPE is going to focus on on-boarding partners who already have “passion and expertise” in the AI ecosystem, said Shattuck. “That’s where we think we are going to find some really complimentary relationships to on board partners into this ecosystem,” he said.

With HPE dramatically simplifying the installation of the platform, the focus is on using AI to drive data insights that solve business problems for customers, said Shattuck.

A Pre-Configured Installer That Reduces Installation Times From Days To Less Than 30 Minutes

The new HPE Ezmeral Software comes with a pre-configured installer for the SaaS data fabric that reduces installation times from “days to less than 30 minutes,” said Shattuck.

In the past when a customer was looking to get up and running on an AI based Ezmeral solution it would require dedicated professional services from HPE, said Shattuck.

“We would spend days, if not weeks, trying to get the software installed running through issues around their DNS and their network and their user authentication, all these things,” said Shattuck. “It could take weeks just to get the environment up and running just so they could play with it. We were out in Israel recently working with a customer and it took just 28 minutes to get it up and running with access to the tools.”

The new installer reduces to one fifth the “environmentals” needed to get the product up and running, said Shattuck.

The new offering also makes it easier to deploy in a public cloud environment, said Shattuck. “If you want to deploy it in one of your cloud service providers put your cloud credentials into the installer and it will just install it into your cloud instance,” he said. “So you don’t even need physical infrastructure. So out of the box it supports hybrid and multi-cloud.”

Consumption Billing For AI Workloads And Unified Analytics

HPE Ezmeral Software has a new consumption based subscription model that is “extremely, extremely price-competitive with the public clouds,” said Shattuck.

The new simplified Ezmeral platform is aimed at providing “predictable and transparent economics” that effectively eliminate the “unpredictable costs and risk of vendor lock” of public cloud environments, said HPE.

The new release of HPE Ezmeral Software effectively reduces the number of software products or tools it takes to produce AI and machine learning based applications in a public cloud or multi-cloud environment, said Shattuck.

In fact, Shattuck said, the new release of Ezmeral has effectively reduced by one third to one half the number of tools a customer would have to subscribe to from a public cloud provider in order to develop and deploy AI based applications.

HPE Ezmeral’s multi-cloud capabities means the number of tools it takes to run and deploy multi-cloud AI based applications is reduced even further, said Shattuck. That’s because there will be a “commonality” of tools with the Ezmeral edge to cloud data fabric across on-premise and multiple cloud environments, he said.

“We see the number of tools that people are going to have to adopt dramatically reduced and with that the corresponding associated price cost,” he said.

HPE Ezmeral at the data fabric level provides file, object, real-time streaming, and database access at a single monthly subscription price, said Shattuck. “We think it is going to be significantly less expensive than if you had bought all of those foundational data services from one of the (public) cloud providers,” he said.

If that isn’t enough, HPE is also providing within its unified analytics software offering “all the tool sets for data engineering, data analysts, some visualization, and machine learning operationalization,” said Shattuck. “That is again another four or five tools you would have to acquire within one of the ([public cloud) providers.”

A New Simplified Unified Analytics Platform

The new HPE Ezmeral Software AI and ML makeover includes a new Unified Analytics SaaS product with an open source based extensible platform for AI applications from AI models to AI development and deployment.

The new Unified Analytics platform supports the wide range of open source products including Kubeflow, Apache Spark, Feast, and other offerings.

The most common open source use tool sets are available out of the box with the new unified analytics platform, said Shattuck.

The new platform is also open and extensible so if a customer wants to work with an emerging open source capability like Dask – an open source parallel computing platform- it can be easily be added to the HPE Ezmeral Unified Analytics platform, said Shattuck.

“Whatever the stable of open source frameworks is we’ll deploy it,” pledged Shattuck. “But customers are going to be able to onboard their own as well and it will be fully integrated.”

That open source support means the common services deployed on the platform like single sign on, networking, security and even Kubernetes management will be provided by HPE, said Shattuck.

“We will manage Kubernetes behind the scenes,” said Shattuck. “So our partners aren’t going to have to have expertise around Kubernetes. Our clients are not going to have to reach out to their IT folks or their Kubernetes admins (administrators). We are going to do all that behind the scenes. No one is going to have to lift a finger in terms of managing the Kubernetes and the plumbing. All they have to do is run the installer, point it to the CPUs, and all of those tools sets are up and running. It is going to be really easy to deploy.”

A Cloud Like Ease Of Use For Hybrid Multi-cloud

The new HPE Ezmeral Software provides a “cloud-like, ease of use experience for hybrid multi-cloud environments,” said Shattuck.

The new HPE Ezmeral Software provides a common set of data services and analytics for complex multi-cloud AI and machine learning workloads from the likes of AWS, Azure, Google Cloud and even popular offerings like Snowflake or Databricks, said Shattuck.

HPE Ezmeral provides “commonality for your data services and analytics whether it is on prem or in anybody’s cloud,” said Shattuck. “So the same ease of use, the same cloud like experience, regardless of whether it is on prem or in the cloud or hybrid.”

The HPE Ezmeral multi-cloud capabilities open the door for customers to move AI and ML workloads between public clouds and on-premise environments, said Shattuck.

Since the data is all connected with HPE Ezmeral Software’s data fabric and application development and deployment is integrated with Unified Analytics, HPE does not “care if you do it in the cloud or on prem,” said Shattuck. “We want to tie all those environments together.”

HPE Does Not See OpenShift As A Competitor Anymore

The new HPE Ezmeral software release no longer takes aim as Red Hat OpenShift as a competitor, said Shattuck.

“We don’t see OpenShift as a competitor anymore,” he said. “When you looked at our old run-time environment that was really heavily going after containerized workloads you could make the argument that we were kind of going head to head with them and really we are trying to pivot away from that.”

HPE now recognizes that as a “generic containerized technology” OpenShift and to a certain degree Tanzu from VMware “own that general container market” said Shattuck. “We don’t want to fight head to head with them. We embed Kubernetes and container management within Unified Analytics and don’t expose it to the world. “

That said, HPE has Kubernetes embedded in Unified Analytics., said Shattuck.

One of HPE’s public road map offerings is support for third party containerization across its portfolio, said Shattuck. So HPE will eventually be able to leverage the container management of OpenShift.

“We’ll put Unified Analytics on top of OpenShift containers and use (HPE Ezmeral Software) data fabric as the common data source but we can use OpenShift for the container management,” he said.

An AI Platform With More Opportunities For HPE Partners

The new HPE Ezmeral Software platform opens the door for more partners to get into the fast growing AI and machine learning solutions marketplace, said Shattuck.

“We’re making it easy for partners to be relevant in this space with an easy to deploy, easy to manage solution that they can bring to their clients,” he said. “It is not going to take heavy services engagement. It is not going to take heavy training on their part to learn it.”

The new simplified HPE Ezmeral Software platform enables partners to help build and deploy AI and ML solutions tailored to a business, said Shattuck. “That is where our channel partners are getting really excited about being able to drive value for their clients by having either horizontal or vertical solutions, leveraging the data and AI toolsets,” he said. “That is where their expertise is. They are not interested in selling IT infrastructure or even in setting up the software. They want to go deploy fraud risk management solutions or network analytics solutions or predictive maintenance solutions for their clients. We are going to make it a lot easier for channel partners to deploy those kinds of solutions.”

The new HPE Ezmeral Software also opens the door for partners to drive consumption based as a service solutions, said Shattuck. “All of our products are going to be available with a consumption based model moving forward,” he said. “So all of our channel partners will then inherit that functionality as well to really compete against (public) cloud and maybe claw back some of that cloud spend.”

Ultimately, the new platform allows partners to compete against public cloud providers and be more relevant in the AI and ML solutions marketplace, said Shattuck. That translates into helping customers “connect to their data and drive AI and ML outcomes,” he said.

Hitting The Market With The AI Explosion In Full Swing

The new robust HPE Ezmeral AI and ML capabilities are hitting the market with the AI explosion in full-force as a result of the ChatGPT AI frenzy, said Shattuck.

“All of a sudden AI has just exploded,” he said. “People who didn’t even know really what AI is are talking about things like ChatGPT. What a lot of people don’t realize is the ChatGPT people are talking about is ChatGPT 4.0. It is not ChatGPT 1.0. There have been three other iterations of ChatGPT.”

The generative AI tool sets that are going “gangbusters” right now have opened people’s minds to the “art of what is possible,” said Shattuck. “But folks are not really sure yet how do they leverage a ChatGPT to make their business more effective and competitive and they are struggling. This is a great opportunity for us and our partners to help customers make that easy.”

HPE teaming with partners are stepping into the AI fray to help customers “leverage the data they have” and put that data into compelling AI based applications for the business, said Shattuck.

“You can have the most sophisticated AI tool in the world, but what makes AI tools and ML algorithms valuable is the data and the customers have the data,” he said. “We can help them get their arms wrapped around the data and we can help them actually deploy AI projects.”

The bottom line: HPE and channel partners together are now poised to help customers “figure out how to leverage AI for the betterment” of their business, said Shattuck.

HPE’s legacy as a partner first organization is going to be critical in helping customers build compelling AI and ML applications, said Shattuck.“HPE has always been one of the most partner friendly tech vendors in the industry,” he said.

“The only way we can scale in the (AI) marketplace is through the partner ecosystem,” he said. “It is the only way that we are going to gain scale. We in the Ezmeral business unit are extremely partner focused and partner friendly. We have got Ezmeral specific partner development managers in every geo (geography) globally to work with our channel partners to accelerate their ability to deploy the software and solutions for our customer.”