Weka and HPE Joint Development Validated Solution for Containerized Workloads

WekaIO, Inc. developed a performance solution with Hewlett Packard Enterprise Development LP (HPE) for demanding AI and ML workloads.

Click to enlarge

The validated solution is available on the HPE Ezmeral Marketplace.

The company offers a performance data platform that was built to address the challenges posed by modern applications that leverage cloud, GPU compute through innovations in how flash is used. Built on WekaFS, it offers the simplicity of NAS, the performance of SAN or DAS and the scalability of object storage. No more compromises between simplicity, speed or scale. With its software-defined architecture, customers can run on-premises, natively in the cloud, and orchestrate their data between the two. The ability to mix and match flash and disk offers customers the best economics. This Ezmeral validated Weka solution is in addition to the existing ProLiant validated solutions currently available from HPE at HPE Solutions for Weka.

HPE Ezmeral Runtime Enterprise is a Kubernetes platform supporting both cloud native and non-cloud native applications. It supports hybrid deployments with both on-premises and public cloud support at the edge, core and cloud for a range of use cases like autonomous vehicles, single instance databases and CI/CD pipelines. Weka, HPE Ezmeral Runtime Enterprise and Ezmeral MLOps enable customers to overcome the challenges of Kubernetes and AI/ML data pipelines and deliver value from the data center to the cloud to the edge (aggregation) with at-scale performance.

“Weka provides our customers with a high performance, SDS platform to support the end-to-end AL/MLOPs data pipeline,” said Matt Hausmann, Ezmeral marketing and GTM, HPE. “HPE Ezmeral and Weka’s file storage platform offer customers the ability to build an application and deploy it anywhere with enterprise security, performance and reliability.“

While containers have become popular because of their lightweight nature, their immutability and portability, and their agility with orchestration provided by Kubernetes, persistent storage, data management and DR remain top barriers to adoption. Kubernetes at scale can be difficult. A validated solution like the HPE Ezmeral and Weka offering provides enterprises with benefits including faster time to market, operational agility, scale and ease of deployment.

As part of this latest integration, the company joins HPE’s growing ecosystem of partners that helps bring speed and agility to the application development lifecycle. The Ezmeral Marketplace leverages the open-source Kubernetes ecosystem and its ISV partners to provide faster time to value. Together, they make it easier for clients to adopt integrated solutions that combine Ezmeral software with validated industry-leading, third-party, commercial, and open-source applications. Weka can be found on the HPE Ezmeral Marketplace.

“Weka sees a growing need to deploy stateful container applications on high-performance storage,” said Shailesh Manjrekar, head, and strategic alliances, WekaIO. “Weka is excited about partnering with HPE to enable solutions with Weka’s scalable parallel file system on HPE Ezmeral Runtime Enterprise. This joint solution is optimized for NVMe and the hybrid cloud, targeting the most-demanding storage, AI/ML and data-intensive technical computing workloads.“

The company‘s software-based, cloud-native data platform allows customers to leverage low-cost S3 storage with fast NVMe flash storage, on-premises or in the public cloud, to build out a massively scalable foundation for enterprise AI projects that attain performance and economics. The firm‘s patented zero-copy architecture, driven by high-performance, scalable multi-protocol support for all file and object, means that customers don’t have to buy, manage, and administer several distinct storage systems across ingest, processing and data protection. Instead, they can leverage the company to integrate flash, HDDs and cloud S3 storage in a single data pool that scales to hundreds of petabytes in a single namespace.

Resource:
HPE blog: High-performance Kubernetes workloads with HPE and Weka