Hammerspace addresses the complexities of modern data management by offering a unified, cloud-based solution. In this show, David Flynn, Co-Founder and CEO at Hammerspace, discusses how the company is solving the challenges of GPUs in data centers, the growing importance of data, and how they are helping organizations manage the increasing complexities of data. He says that their data platform enables seamless data transfer between various storage systems and data centers, delivering it in massive, parallel quantities with unparalleled performance, all through a standards-based approach.

The challenges of GPUs in data centers and how Hammerspace is solving them

• Hammerspace is a data platform that streamlines data across various storage systems and data centers, providing global access and high performance by integrating data orchestration, global file systems, and hyperscale Network Attached Storage (NAS).
• Flynn notes that GPUs, once used mainly for media and entertainment, are now essential for AI, managing large data volumes for training and inference.
• Flynn highlights the challenges with GPUs including high power consumption and the need for efficient cloud-based solutions to manage bursty workloads. He explains how Hammerspace is solving these challenges.

How Hammerspace helps address the challenges of managing large volumes of data

• Hammerspace addresses the challenges of managing large volumes of data and bursty workloads by enabling granular, on-demand data movement at the file level, rather than handling large data pools monolithically.
• Flynn discusses how Hammerspace’s approach is crucial for efficiently accessing GPUs, optimizing storage, and reducing compute costs, particularly in the context of AI model training.
• The platform utilizes principles similar to containerized microservices orchestration to move data efficiently across different data centers, which helps optimize egress and minimize bandwidth costs.

The scope of use cases and industries Hammerspace supports

• Hammerspace is utilized by major organizations like Blue Origin, Meta, and Los Alamos National Labs to optimize data management across cloud environments, enhance AI training efficiency, and support supercomputing.
• Flynn talks about how Hammerspace serves diverse industries and is vital for handling data as AI transitions to machine-driven processing.
• Hammerspace’s new support for the S3 protocol, enabling data to be accessed both as a file system and as objects in an object store. Flynn highlights how this enhancement facilitates handling distributed or edge use cases.

The growing importance of data and the need for new approaches to data management

• Flynn discusses how data is becoming crucial across various sectors, even in traditionally physical industries like energy exploration and supply chain management.
• Flynn emphasizes that data, particularly unstructured data, has grown in importance as AI enables its analysis and use. This shift to data-driven operations requires new approaches to data management
• Efficient data orchestration is now essential for handling the complexities of modern data usage.

What differentiates Hammerspace from its competitors?

• Flynn explains that, unlike traditional point solutions such as cloud gateways or backup and restore systems, Hammerspace offers a comprehensive, universal solution for data management.
• Flynn highlights Hammerspace’s edge over competitors with its comprehensive, universal data management platform, contrasting it with competitors’ more specialized or fragmented tools.

How Hammerspace simplifies data governance and compliance

• Flynn highlights that Hammerspace simplifies data governance and compliance by unifying data within a single namespace, unlike traditional methods that complicate compliance. Its single-pane view and policy-based data movement improve security and management.
• Flynn also emphasizes Hammerspace’s advantage in performance by integrating with standard operating systems rather than requiring complex, proprietary systems, providing high-speed access without the need for specialized maintenance.
• Flynn advocates for a unified, cloud-based approach to maximize GPU usage and ensure regulatory compliance, highlighting the essential role of contemporary data management solutions.
• Hammerspace is envisioned as a public highway system for data storage, offering improved speed and efficiency while being compatible with existing systems.
• The company’s CTO, who maintains the NFS client stack kernel, has been instrumental in developing Hammerspace’s advanced data management system, contributing to its enhanced performance.

Guest: David Flynn (LinkedIn)
Company: Hammerspace (Twitter)
Show: Let’s Talk

This summary was written by Emily Nicholls.