Hammerspace for

High Performance Computing

DATA SHEET

HPC organizations need to design data architectures to ingest data at the edge, facilitate high-performance data creation, leverage cloud resources, and make data available to remote teams.

Hammerspace delivers high-performance data storage and data orchestration in a single global parallel file system. This gives direct access to all data created, in any location, on any storage system making it possible to feed data hungry applications and researchers with the information they need to derive insights and make scientific discoveries.

The Hammerspace file system allows organizations to take full advantage of the performance capabilities of any server, storage system, and network anywhere in the world. This capability enables a unified, fast, and efficient global data environment for the entire workflow, from data creation to processing, collaboration, and archiving across edge devices, data centers, and public and private clouds.

  • High-Performance Across Data Centers
    and to the Cloud: Saturate the Available
    Internet or Private Links Available
  • High-Performance Across Interconnect
    within the Data Center: Saturate Ethernet
    or InfiniBand Networks within the Data
    Center
  • High-Performance Server-local IO:
    Deliver to Applications Near Theoretical
    IO Subsystem Maximum Performance of
    Cloud Instances, VMs, and Bare
    Metal Servers

Hammerspace Will Never Be The Bottleneck

For years, HPC teams have had no option but to use the technologies designed for single data center deployments to piece together their edge and hybrid cloud data architectures. They have been forced to add additional software layers for data orchestration that either create data path friction, making high-performance data access across decentralized environments impossible.

The Hammerspace Global Data Environment was built from the ground up to take full advantage of any hardware infrastructure it runs on. At last, a parallel file system exists that creates a single, high-performance data namespace that unifies all data across edge, HPC centers, and cloud!

Hammerspace enables HPC organizations to run their entire workflow in a single, global parallel system. With Hammerspace:

1. Capture/ingest data as fast as modern instruments and high-performance compute clusters can create the data
2. Intelligently move data to ML and AI applications for analysis
3. Provide high-performance, local read/write access to shared data for remote collaboration
4. Archive, without creating unwanted duplicate copies, on the appropriate storage tier

High Performance Starts With Metadata

HPC architects no longer need to select different software solutions for their high-performance file system and for their data management and data orchestration. Thanks to the magic of the Hammerspace architecture, all users and applications interact with extremely high-performance shared metadata to ensure fast visibility to all data across the entire Global Data Environment.

High-performance storage environments now share the same metadata with edge, cloud, and archive data, making finding and using data simple.

Rich, actionable metadata is critical to deriving value from data. Hammerspace enables custom metadata tagging at the users’ desktop and automatically through application integration with the Hammerspace UI, CLI, and open API options.

A logical view of the Hammerspace Architecture. Hammerspace software is deployed as a fully integrated software solution built upon open standards, with a single installer that includes all components necessary for deployment on bare metal servers, VMs or in cloud-based compute environ- ments. No external software dependencies are required, including the Linux OS.

The Hammerspace metadata not only ensures high-performance data access, it also is the intelligence used by data objectives that place data on the correct storage, when created or when accessed, based upon workflows and/or business-rules. In this way it automatically aligns data placement to storage with the correct performance profile for the job, without interrupting users or adding complexity for IT staff.

“Data being consumed by both traditional HPC modeling and simulation workloads, and modern AI and HPDA workloads, is being generated, stored, and shared between a disparate range of resources, such as the edge, HPC data centers, and the cloud,” said Mark Nossokoff, Research Director at Hyperion Research. “Current HPC architectures are struggling to keep up with the challenges presented by such a distribute d data environment. By addressing the key areas of collaboration at scale while supporting system performancecapabilities and minimizing potential costly data movement in HPC cloud environments, Hammerspace aims to deliver a key missing ingredient that many HPC users and system architects are looking for.”

Hammerspace is the Solution that Decentralized High-Performance Data Organizations Have Been Asking For

    • Finally, you can manage your entire file data global workload in a single Global Data Environment. With Hammerspace, IT administrators can use a single solution to bridge existing storage silos to enable global data access, administration, and protection across multiple vendor storage types and locations.
    • Data security is ensured by eliminating unwanted copy proliferation and implementing automated service-level objectives based upon business rules to globally manage and protect all data across any storage type and location for the entire organization.
    • Data is globally available to any authorized user or application, on any infrastructure and in any location, anywhere in the world.

“I see high-performance access to global data as the next step in innovation for high-performance environments. The challenge of fast networks and fast computers has been well-solved for years but making remote data available to these environments was a poorly solved problem until Hammerspace came into the market. Hammerspace makes it possible to take cloud and data utilization to the next level of decentralization, where data resides,” said Eyal Waldman, co-founder and previous CEO of Mellanox Technologies, and Hammerspace Advisory Board Member.

Finally, you can manage your entire file data global workload in a single Global Data Environment.