HPE GreenLake edge-to-cloud platform rolls out industry’s first cloud-native unified analytics and data lakehouse cloud services optimized for hybrid environments
SEPTEMBER 21, 2021 • BLOG POST • VISHAL LALL, SVP & GM, HPE GREENLAKE CLOUD SERVICES
IN THIS ARTICLE
- First cloud-native solution to bring Kubernetes-based Apache Spark analytics and the simplicity of unified data lakehouses using Delta Lake on-premises
- Only data fabric to combine S3-native object store, files, streams and databases in one scalable data platform
- Cloud-native unified analytics platform enables customers to modernize legacy data lakes and warehouses without complex data migration, application rewrites or lock-in
- 37 solution partners support HPE Ezmeral with 15 joining the HPE Ezmeral Partner Program in the past 60 days
Built on HPE Ezmeral software, analytics and data science teams benefit from frictionless access to data from edge to cloud and a unified platform for accelerated Apache Spark and SQL
In the Age of Insight, data has become the heart of every digital transformation initiative in every industry, and data analytics has become critical to building successful enterprises. Simply put, data drives competitive advantage. However, for most organizations, significant challenges remain for organizations to successfully execute data-first modernization initiatives. Until now, organizations have been stuck with legacy analytics platforms that were either built for a pre-cloud era and lack cloud-native capabilities, or require complex migrations to public clouds, risking vendor lock-in, high costs and forcing adoption of new processes. This situation has left the big data and analytics software market1 -- which IDC forecasts will reach $110 billion by 2023 – ripe for disruption.
Today, I am excited to announce two disruptive HPE GreenLake cloud services that will enable customers to overcome these trade-offs. There are four big value propositions we optimized for:
1. Seamless experience for a variety of analytics, SQL, and data science users
2. Top-notch performance
3. Choice and open ecosystem by leveraging pure open source in a hybrid environment
4. An intense focus on reducing TCO by up to 35% for many of the Workloads we are targeting
Built from the ground up to be open and cloud-native, our new HPE GreenLake for analytics cloud services will help enterprises unify, modernize, and analyze all of their data, from edge-to-cloud, in any and every place it's stored. Now analytics and data science teams can leverage the industry’s first cloud-native solution on-premises, scale up Apache Spark lakehouses, and speed up AI and ML workflows. Today’s news is part of a significant set of new cloud services for the HPE GreenLake edge-to-cloud platform, announced today in a virtual launch event from HPE. The new HPE GreenLake for analytics cloud services include the following:
HPE Ezmeral Unified Analytics
HPE now offers an alternative to customers previously limited to solutions in a hyperscale environment by delivering modern analytics on-premises, enabling up to 35%2 more cost efficiencies than the public cloud for data-intensive, long running jobs typical in mission critical environments. Available on the HPE GreenLake edge-to-cloud platform, HPE Ezmeral Unified Analytics is the industry’s first unified, modern, hybrid analytics and data lakehouse platform.
We believe it is the first solution to architecturally optimize and leverage three key advancements simultaneously which no one else in the industry has done.
1. Optimize for a Kubernetes based Spark environment for on-premises deployment providing the cloud-native elasticity and agility customers want
2. Handle the diversity of data types from files, tables, streams, and objects in one consistent platform to avoid silos and make data engineering easier
3. Embrace the edge by enabling a data platform environment which can span from edge to hybrid cloud
Instead of requiring all of your data to live in a public cloud, HPE Ezmeral Unified Analytics is optimized for on-premises and hybrid deployments, and uses open source software to ensure as-needed data portability. We designed our solution with the flexibility and scale to accommodate enterprises’ large data sets, or lakehouses, so customers have the elasticity they need for advanced analytics, everywhere.
Just a few key advantages of HPE Ezmeral Unified Analytics include:
- Dramatic performance acceleration: Together NVIDIA RAPIDS Accelerator for Apache Spark and HPE Ezmeral can accelerate Spark data prep, model training, and visualization by up to 29x3, allowing data scientists and engineers to build, develop, and deploy at scale analytics solutions into production faster.
- Next-generation architecture: We have built on Kubernetes and added value through an orchestration plane to make it easy to get the scale-out elasticity customers want. Our multi-tenant Kubernetes environment supports a compute-storage separation cloud model, providing the combined performance and elasticity required for advanced analytics, while enabling users to create unified real-time and batch analytics lakehouses with Delta Lake integration.
- Optimized for data analytics: Enterprises can create a unified data repository for use by data scientists, developers, and analysts, including usage and sharing controls, creating the foundation for a silo-free digital transformation that scales with the business as it grows, and reaches new data sources. Support for NVIDIA Multi-Instance GPU technology enables enterprises to support a variety of workload requirements and maximize efficiency with up to seven instances per GPU.
- Enhanced collaboration: Integrated workflows from analytics to ML/AI span hybrid clouds and edge locations, including native open-source integrations with Airflow, ML Flow, and Kubeflow technologies to help data science, data engineering, and data analytics teams collaborate and deploy models faster.
- Choice and no vendor lock-in: On-premises Apache Spark workloads offer the freedom to choose deployment environments, tools, and partners needed to innovate faster
“Today’s news provides the market with more choice in deploying their modern analytics initiatives with a hybrid-native solution, enabling faster access to data, edge to cloud,” said Carl Olofson, Research Vice President, IDC. “HPE Ezmeral is advancing the data analytics market with continued innovations that fill a gap in the market for an on-premises unified analytics platform, helping enterprises unlock insights to outperform the competition.”
HPE Ezmeral Data Fabric Object Store
Our second disruptive new solution is the HPE Ezmeral Data Fabric Object Store: the industry’s first Data Fabric to combine S3-native object store, files, streams and databases in one scalable data platform that spans edge-to-cloud. Available on bare metal and Kubernetes-native deployments, HPE Ezmeral Data Fabric Object Store provides a global view of an enterprise’s dispersed data assets and unified access to all data within a cloud-native model, securely accessible to the most demanding data engineering, data analytics, and data science applications. Designed with native S3 API, and optimized for advanced analytics, HPE Ezmeral Data Fabric Object Store enables customers to orchestrate both apps and data in a single control plane, while delivering the best price for outstanding performance.
We are proud of the innovation that has resulted in what we believe is an industry first: A consistent data platform which is able to handle a diversity of data types, is optimized for analytics, and is able to span from edge to cloud.
Several key features include:
- Optimized performance for analytics: Designed for scalable object stores, HPE Ezmeral Object Store is the industry's only solution that supports file, streams, database, and now object data types within a common persistent store, optimized for best performance across edge-to-cloud analytics workloads.
- Globally synchronized edge-to cloud data: Clusters and data are orchestrated together to support dispersed edge operations, and a single Global Namespace provides simplified access to edge-to-cloud topologies from any application or interface. While data can be mirrored, snapshotted, and replicated, advanced security and policies ensure the right people and applications have access to the right data, when they need it.
- Continuous scaling: Enterprises can grow as needed by adding nodes and configuring policies for data persistence while the data store handles the rest.
- Performance and cost balance: Adapting to small or large objects, auto-tiering policies automatically move data from high-performance storage to low-cost storage.
Expanding the HPE Ezmeral Partner Ecosystem
We first introduced the HPE Ezmeral Partner Program in March 2021, enabling the rapid creation of streamlined, customized analytics engines and environments based on full stack solutions validated by trusted ISV partners. With 76% of enterprises expecting to be using on-premises, third-party-managed private cloud infrastructure for data and analytics workloads within the next year4, we're excited to announce six new ISV partners today, including: NVIDIA NGC, Pepperdata, Confluent, Weka, Ahana and gopaddle.
“NVIDIA’s contributions to Apache Spark enable enterprises to process data orders of magnitude faster while significantly lowering infrastructure costs,” said Manuvir Das, head of Enterprise Computing, NVIDIA. “Integrating the NVIDIA RAPIDS Accelerator for Apache Spark and NVIDIA Triton Inference Server into the HPE Ezmeral Unified Analytics Platform streamlines the development and deployment of high-performance analytics, helping customers gain immediate results at lower costs.”
“Today, companies are using Spark to build their high-performance data applications, accelerating tens to thousands of terabytes of data transitioning from data lakes to AI data modeling,” said Joel Stewart, Vice President Customer Success, Pepperdata. “Pepperdata on HPE Ezmeral Runtime Enterprise can help reduce operating costs and provide deep insights into their Spark applications to improve performance and reliability.”
Since the HPE Ezmeral Partner Program launched, we've added 37 solution partners5 to support our customers' core use cases and workloads, including big data and AI/ML use cases. The Partner Program is also adding support today for open-source projects such as Apache Spark, offering enterprises the ability to transition workloads to a modern, cloud-native architecture.
HPE GreenLake edge to-cloud platform and HPE Ezmeral are transforming enterprises – and HPE
As an important component of HPE GreenLake cloud services, the HPE Ezmeral software portfolio help enterprises such as GM Financial and Bidtellect advance modern data analytics initiatives. Since it was first introduced in June 2020, HPE Ezmeral has secured dozens of new customers, with significant competitive wins over both traditional big data players, as well as public cloud vendors.
Since vast volumes of applications and data remain will remain on-premises and at the edge as enterprises continue their digital transformations, our elastic, unified analytics solutions will help customers extract maximum value from their data, wherever it lives and moves, from edge-to-cloud. We look forward to working with you to make the most of your data as the Age of Insight continues to reshape enterprises around the world.
Availability and Additional Resources
HPE Ezmeral Unified Analytics and HPE Ezmeral Data Fabric Object Store will be available as HPE GreenLake cloud services beginning November 2021 and Q1 2022, respectively.
Learn more about today’s news from the experts. Join these deep dive sessions as I chat with:
- Keith White, SVP & GM, HPE GreenLake Cloud Services Commercial Business on how enterprises are accelerating transformation for greater business outcomes.
- Matt Maccaux, Global Field CTO, on next generation analytics and data services.
HPE and the HPE logo are trademarks or registered trademarks of HPE and/or its affiliates in the U.S. and other countries. Third-party trademarks mentioned are the property of their respective owners.
1 IDC, Worldwide Big Data and Analytics Software Forecast, 2021–2025, July 2021
2 Based on internal HPE competitive analysis, September 2021
3 Technical Paper: HPE Ezmeral for Apache Spark with NVIDIA GPU, published September 2021
4 451 Research Voice of the Enterprise: Data & Analytics, Data Platforms 2021
5 Internal HPE documentation on list of partners maintained by the group