HPE has given its HPE Ezmaral Software platform a new update for more straightforward execution of ML and AI projects. These include open-source tooling, object and streaming data source integration, and greater interoperability with multiple cloud environments.
The latest version of the HPE Ezmaral Software platform for data analytics should end the chaos that all analytics tools and workloads bring around performance, cost and compliance requirements. It also addresses issues around data control, potential vendor and data lock-in and performance around public clouds, as well as issues posed by homegrown open-source tools.
In the new version, the HPE Ezmeral Data Fabric Software provides hassle-free access to multiple data sources and formats. These include files, objects, tables and data streams. Among other things, users get highly detailed monitoring tools and automated policy management for workload optimization. This is based on performance, data location, sovereignty, cost and compliance requirements. This software is now available for the first time as a SaaS solution.
The HPE Ezmeral Unified Analytics Software suite offers several “as-a-service” open-source tools. These include the Apache Airflow workflow management platform, the Spark analytics framework, the Superset visualization platform, the Presto SQL distributed data store and the Ray distributed computing framework. Also available are the Feast store for ML, the Kubeflow platform for Kubernetes and the ML management platform MFlow.
For training and deploying AI applications, it offers several new connectors with Snowflake, MySQL, DeltaLake, Teradata and Oracle, among others.
The new HE Ezmaral Software platform is specifically designed for hybrid environments and usable in edge, colocation, on premises and in public cloud environments. This should save operating costs for customers because it performs analytics where both data and compute power are located.
Also read: HPE brings GreenLake cloud experience to on-prem block and file storage