Search
Close this search box.

VAST Data Unveils New Data Center Architecture to Accelerate AI

VAST Data Unveils New Data Center Architecture to Accelerate AI

The News: VAST Data unveiled a series of announcements, all of which advance VAST’s presence in the AI market. The announcements include a new architecture based on NVIDIA BlueField DPUs, a partnership with SuperMicro, and new capabilities focused on write performance. More information can be found on the VAST Data website.

VAST Data Unveils New Data Center Architecture to Accelerate AI

Analyst Take: VAST Data has recently made a series of announcements furthering its push into the AI space, including a new architecture designed around NVIDIA BlueField DPUs. The new architecture, which was codesigned with NVIDIA, allows for stateless VAST containers to be run on NVIDIA BlueField DPUs embedded in GPU servers. This approach aims to remove layers of compute and networking from a typical deployment, while integrating storage and database services directly into AI servers. VAST has stated that the new architecture has been deployed and is being tested by CoreWeave.

The architecture delivers a new way to deploy VAST software for AI use cases by leveraging NVIDIA BlueField DPUs. The architecture offers an efficient and scalable way to deploy VAST in large AI and GPU environments, while reducing the cost, power, and overhead of a typical deployment. The reduction of power is especially notable due to ongoing sustainability concerns regarding the power and resource requirements of AI.

Alongside unveiling the new architecture, VAST has released two other notable announcements, including a partnership with SuperMicro and new write performance features. The partnership with SuperMicro combines the VAST Data Platform with NVIDIA-certified systems from SuperMicro. VAST states that the collaboration provides “a full-stack, end-to-end AI solution aimed at simplifying the creation and expansion of large-scale AI deployments.” Ultimately, the partnership provides another option for organizations to deploy VAST for their AI requirements.

The third notable announcement, in a blog post from VAST cofounder Jeff Denworth, concerns features to improve write performance. The focus on improving write performance stems from shifting read/write patterns in AI workloads, due to the need to frequently write large checkpoints. The new features include SCM RAID and Spillover. VAST claims that the switch from SCM Mirroring to SCM RAID will provide a 50% performance increase on writes. The Spillover feature, which will be available in version 5.2, allows large checkpoint writes to write directly to QLC flash. While writing directly to QLC may cause some concerns about degradation, VAST has notably stated that the new feature “will not jeopardize our ability to ensure a decade of QLC SSD endurance”.

All three announcements demonstrate positive momentum from VAST Data and showcase acceleration of the company’s AI vision. While seemingly separate announcements, all three announcements further VAST’s push into the AI market, whether through new architectures for large-scale GPU deployments, new partnerships to deploy end-to-end AI solutions, or new features to improve write performance for AI workloads.

Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.

Other Insights from The Futurum Group:

Demystifying AI, ML, and Machine Learning

VAST Data Announces Partnership with Genesis Cloud

Weka Partners with NexGen Cloud to Build Sustainable AI Supercloud

Author Information

Mitch comes to The Futurum Group through the acquisition of the Evaluator Group and is focused on the fast-paced and rapidly evolving areas of cloud computing and data storage. Mitch joined Evaluator Group in 2019 as a Research Associate covering numerous storage technologies and emerging IT trends.

With a passion for all things tech, Mitch brings deep technical knowledge and insight to The Futurum Group’s research by highlighting the latest in data center and information management solutions. Mitch’s coverage has spanned topics including primary and secondary storage, private and public clouds, networking fabrics, and more. With ever changing data technologies and rapidly emerging trends in today’s digital world, Mitch provides valuable insights into the IT landscape for enterprises, IT professionals, and technology enthusiasts alike.

SHARE:

Latest Insights:

Managing Cloud Costs Amid AI and Cloud-Native Adoption
Paul Nashawaty and Steven Dickens of The Futurum Group cover IBM's acquisition of Kubecost, a cost monitoring and management tool for Kubernetes, marking a step toward providing a comprehensive cost management platform for cloud-native applications.
Veeam Makes a Strategic Move to Enhance Positioning in Next-Generation, AI-Driven Cyber Resilience
Krista Case, Research Director at The Futurum Group, covers Veeam’s acquisition of Alcion and its appointment of Niraj Tolia as CTO. The move will strengthen its AI cyber resilience capabilities.
Google’s New Vault Offering Enhances Its Cloud Backup Services, Addressing Compliance, Scalability, and Disaster Recovery
Krista Case, Research Director at The Futurum Group, offers insights on Google Cloud’s new vault offering and how this strategic move enhances data protection, compliance, and cyber recovery, positioning Google against competitors such as AWS and Azure.
Capabilities Focus on Helping Customers Execute Tasks and Surface Timely Insights
Keith Kirkpatrick, Research Director with The Futurum Group, shares his insights on Oracle’s Fusion Applications innovations announced at CloudWorld, and discusses the company’s key challenges.