HPE Ezmeral: Uncut

Future-proof your data and analytics investment with HPE Ezmeral

HPE-Ezmeral-AMP-Blog2.jpgThe data management process is constantly evolving. That’s because technology continues to evolve, businesses grow and change over time, and unforeseen events like the pandemic can force enterprises to reassess their business strategy.

Pave the path for a sustainable future

To ensure business continuity and innovation, it is essential to modernize your big data and analytics platform. Your next generation platform must be able to do three things well from edge to core to cloud:

  • Unify diverse data sets
  • Store, manage, and move data with agility
  • Deploy artificial intelligence (AI) and analytics for real-time insights and decision making

To pave the path toward a sustainable future, you also must make some big decisions.  These include things such as whether to move to the cloud, adopt containers and Kubernetes, or embrace open-source technology. You also need to consider how to ensure the performance, security, and high availability of your infrastructure.

The struggle is real for Cloudera customers

Unfortunately for many Cloudera customers, the path to the future is fraught with uncertainty. Cloudera customers who have invested in Hadoop-based platforms are at a crossroads: they face a costly migration from existing Cloudera and Hortonworks to Cloudera’s Cloudera Data Platform (CDP). And with the End of Support for Cloudera products looming, the CDP essentially has no real future other than sustaining legacy, as it still doesn't allow organizations to fully take advantage of the latest platforms and analytic technologies. 

Moreover, Cloudera customers need to make a strategic decision as to how to protect their investment. They must decide: Do we migrate from Cloudera Distribution Hadoop (CDH) to CDP or another data management platform? What’s our long-term path for a sustainable future? Are we getting enough value from our data? Do we continue fixing our data lake issues? When should we begin the migration, and how should we plan and coordinate, especially if business-critical applications or production jobs are running? Do we have the technical resources on our team to complete the migration?

Customer turns to HPE to provide low-risk, lower cost migration

Recently, HPE was approached by a major financial institution to help them with these difficult questions. The bank operates hundreds of branches, thousands of ATM’s, and over 200,000 POS terminals. They generate an enormous amount of customer and financial data. Consequently, they were facing data management challenges stemming from data siloes and a complex environment made up of multiple solution vendors: Cloudera, Oracle, and Teradata. As a result, their environment was costly and inefficient to run, and they had no way to operationalize machine learning models and apply AI for real-time insights. All of this was hindering their ability to compete with digital disruptors and innovate with speed.

In collaboration with the client, HPE conducted an AMP assessment, a proven best practices migration methodology where HPE data and analytics experts set out to understand the client’s existing platform. During an intense week of collaboration between key stakeholders and HPE experts, the in-depth AMP assessment consists of the following steps.

  1. Analyze the current-state platform
  2. Map the current state findings against the desired business outcomes
  3. Prescribe a systematic plan to achieve the future-state goals.   

Below is a high-level snapshot of the AMP assessment performed for the client.

  • Two appliances powered by Cloudera: Teradata, Oracle BDA
  • About 1500 jobs per day
  • Close to 1 PB of data
  • Ecosystem component and 3rd party applications include Hive, KUDU, Kafka, StreamSets, HUE, Spark, Oozie, Sqoop, Sentry and SolR
  • Encountering challenges with version mismatches (software and library versions)

Desired business outcomes:

  • Provide a data platform for advanced analytics and machine learning
  • Provide a Data Science Sandbox for data exploration and advanced analytics
  • Facilitate data offload and migrations
  • Support application migrations
  • An architecture that provides the customer with a familiar environment, while at the same time provide additional EDF features and improved performance
  • A distributed file system that supports the HDFS protocol
  • Two flavors of NoSQL Databases supporting the HBase API
  • Messaging framework that supports the Kafka API
  • Yarn support
  • Support Distributed processing and query engines
  • Support Common security model
  • Snapshots, local and remote mirroring, and tiered storage volumes
  • A smooth transition, allowing apps and jobs to be migrated slowly in a phased approach
  • A technical foundation that allows the customer to take the next step in their analytics journey, including:
    • Separation of compute and storage
    • Container orchestration on top of Kubernetes
    • Containerization of workloads in support of microservices
    • Large ecosystem of Opensource, Proprietary and Custom Operators such as the Spark operator
  • Machine Learning Operations (MLOps)


At the completion of the AMP assessment, HPE provided a clear, step-by-step implementation plan. This plan including how to migrate from Cloudera to HPE Ezmeral, the hybrid software platform that empowers organizations and all users (data scientists, developers, architects, and IT operations) to leverage containers and Kubernetes to operationalize analytics, AI, machine learning (ML), and data-intensive workloads at scale. In addition, HPE Ezmeral provides the foundational data fabric that delivers enterprise-wide global access to data – from edge to core to cloud. With HPE Ezmeral, the client can run the most advanced and demanding workloads across open-source platforming and tooling (such as Apache Spark on Kubernetes) while avoiding lock-in.

With the migration currently underway, the client looks forward to rolling out their future-ready, next generation platform. The new unified and scalable data platform will empower them to operationalize their machine learning models, deploy AI, and extract the most value from their data – quickly and cost-effectively. This capability ensures they can disrupt their competition as opposed to being disrupted.

“We’re excited to see our client making the switch to HPE Ezmeral. By employing our professional services expertise to ensure best practices are adopted, they can scale and optimize their environment and achieve performance levels that were previously beyond their reach. This means they will be more than a match for current and future SLAs in a very demanding, fast-paced, and highly-regulated sector,” says Robert Yuile, VP, Professional Services EMEA and APJ at HPE.

Developed from an extensive history of successful migrations, the HPE AMP Assessment Program brings the expertise needed to help you migrate onto HPE Ezmeral with no risk and zero downtime. With AMP, Cloudera customers can pave the path for a sustainable future. Additionally, HPE provides cloud services through HPE GreenLake, so that you can benefit from deploying HPE Ezmeral as-a-service for greater agility, lower cost, and flexibility.

Ready to get started or have questions? Please contact ezmeral-services@hpe.com.

About the author:

SteveLeung.jpgSteve Leung

As the WW Managed Services leader for HPE enterprise software solutions, Steve brings deep subject-matter expertise in big data analytics, machine learning, cloud and containerization technologies while leading a team that manages environments for HPE Ezmeral customers.




Hewlett Packard Enterprise


0 Kudos
About the Author