Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


WekaIO Partners with HPE to Develop All-Flash Storage for HPC and AI

Today WekaIO announced a partnership with HPE to deliver integrated flash-based parallel file system capabilities that can significantly accelerate compute-intensive workloads. The WekaIO Matrix software-defined storage solution is validated for deployment within HPE environments – including the HPE Apollo Gen10 System platform that delivers rich capabilities for high-performance computing (HPC), artificial intelligence (AI) and machine learning (ML) use cases.

At HPE we’re committed to providing innovative solutions for our customers in the rapidly growing markets for high-performance computing, artificial intelligence and machine learning,” said Bill Mannel, Vice President and General Manager, HPC and AI Segment Solutions, Hewlett Packard Enterprise. “The combination of WekaIO Matrix with HPE Apollo Systems is an option that enables customers to maximize the throughput of their HPC environment by making it easy to scale storage capacity and performance to new levels without the requirement to modify compute codes or HPC workflows.”

The agreement creates an offering that targets the space, energy, and processor attributes of supercomputing. The HPE portfolio of HPC and AI solutions—including Apollo 2000, Apollo 6000, Apollo 6500, and SGI 8600—all provide rich architectures for leveraging high-performance flash storage both within the actual compute platforms as well as across high-performance interconnect fabrics. WekaIO Matrix delivers high-performance file system technology running on top of the HPE Apollo System.

Key features of the WekaIO solution include:

  • A distributed file system to share data and meet full POSIX compliance
  • Storage density scaling to many petabytes per rack when paired with high-density flash storage devices
  • Disaggregated storage that eliminates the need to retrofit existing servers and enables independent scaling of capacity
  • Native NVMe integration for massive performance gains, lower latency and performance that is faster than traditional filers

WekaIO is delighted to work with HPE to deliver highly-accelerated storage capabilities that can eliminate the potential for I/O bottlenecks that can occur as customers dramatically increase the compute horsepower and fabric bandwidth within their HPC and ML environments,” said Liran Zvibel, co-founder and Chief Technology Officer at WekaIO. “Our all-flash parallel file system scales to meet the demanding IOPS and throughput requirements that are created when customers leverage HPE’s platforms for high-performance data processing. The result is a low latency, throughput-optimized environment capable of matching storage performance to meet the compute and fabric attributes of the solution.”

This technology is ideal for organizations that require the highest possible performance in the smallest footprint, supporting data intensive and performance hungry applications such as electronic design automation, genomic sequencing, video processing and breakthrough areas of scientific discovery.

Amidst an increasing diversity of high-performance workloads, the industry has seen an influx of new technologies, including the widespread adoption of flash storage for higher-throughput data movement,” said industry analyst Addison Snell. “In general, new hardware implementations only rise to their full solution potential when there is a corresponding advancement in software. This philosophy is at the heart of HPE’s integration with WekaIO Matrix.”

Liran Zvibel will be presenting “Next Generation NVMe-Native Parallel File System Accelerating HPC Workloads,” at SC17 on Thursday, November 16, from 9:45am -10:50am in the HPE booth #925.

Sign up for our insideHPC Newsletter

Leave a Comment

*

Resource Links: