SC24: Vdura Data Platform Software

VDURA, Inc. announced its Data Platform software, designed to set new industry benchmarks in performance, scalability, and data management simplicity. This software solution addresses the surging demands of AI and HPC, empowering enterprises and service providers to efficiently manage data, achieve high-speed performance, and accelerate AI initiatives without the traditional trade-offs in reliability and scalability.

The latest Data Platform is a modernization on the previous releases with a move to a fully parallel, microservices-based architecture, a new flash-optimized metadata engine and an enhanced object storage layer. Together, these features ensure unmatched efficiency, scalability, and durability across on-premises, cloud, and hybrid environments.

Software-Driven Infrastructure for Modern AI and HPC
“As AI workloads expand, data infrastructure must support massive write throughput for frequent GPU checkpointing, rapid random reads for RAG-driven inference, billions of files of varying sizes, and substantial capacity for large training datasets-all while ensuring absolute data protection,” said Ken Claffey, CEO. “VDURA’s platform is purpose-built to address these demands with >11 nines of durability, exceptional performance, and uncompromising availability. Our solution is not only powerful and scalable but also easy to manage and economically efficient, giving organizations confidence that their valuable insights are secure and always accessible.”

Key Software Innovations

• VeLO: Velocity Layered Operations – It is a new Key Value Store used within the platforms director layer for handling small files and metadata operations. It supports the diverse range of workloads found in AI. It is optimized for flash and delivers up to two million IO/s per instance. It supports an infinitely scalable number of VeLO instances in the same global namespace.
• VPOD: Virtualized Protected Object Device – They are discrete, virtualized, and protected storage units, forming the foundation of storage in hybrid nodes. These infinitely scalable VPOD instances operate within a unified global namespace, providing high performance and flexible scalability. Data is safeguarded through erasure coding across multiple VPODs in the cluster, with an optional, additional layer of erasure coding within each VPOD for enhanced protection. This multi-layered approach achieves up to 11x of durability with minimal overhead. Data reduction services further optimize efficiency, reducing costs and total ownership expenses.
• Advanced Data Management and Placement – The platform is now deployed as discrete microservices, simplifying its seamless deployment across thousands of nodes, helping ensure linear performance scalability with an infinitely expandable global namespace. The architecture also introduces intelligent data orchestration capabilities that enable optimal data placement across discrete tiers of storage but all within a single unified data and control plane within a single, unified namespace.

Certified Hardware Platforms for Optimized Performance and Flexibility
To complement its powerful software, the company also introduces the V5000 Certified Platform, designed to unlock the potential of its Data Platform. This hardware platform combines modular storage nodes that can be configured for maximum throughput, capacity, or IO/s, depending on the user’s requirements.

Highlights of the V5000 Platform:
1U Director Nodes powered by the latest AMD EPYC 9005 processors, NVIDIA ConnectX-7, Broadcom 200GbE and SAS 4 adapters and Phison PCIe NVMe SSDs, optimized for high IO/s and metadata operations through VeLO instances.

Hybrid Storage Nodes incorporating the same 1U server used with the Director Node and 4U JBODs running VPOD instances for cost-effective bulk storage with high performance and reliability.

Composability for Any Budget
The firm’s platform supports a flexible ratio of HDD to flash device content, allowing enterprises to optimize highly composable configurations for specific cost, performance, and durability needs. This flexibility empowers users to balance capacity and throughput while maintaining a consistent global namespace, making the Data Platform suitable for diverse capacity, price and performance requirements.

Future-Proofed for Continued Innovation
As part of firm’s dedication to continuous improvement, the Data Platform is set to expand further with upcoming releases. In 2025, VDURA will introduce the Cloud Edition, enabling seamless cloud and hybrid deployments. Additionally, a new all-flash platform, designed for ultra-high IO/s, will be available by 2H25, catering to the needs of AI and HPC environments that demand the highest data processing speeds.

VDURA will be at SC24 in Atlanta, GA from November 17-22.

Proven Success Across Industries
“After a thorough evaluation, we selected VDURA’s Data Platform and V5000 for its superior price/performance and durability. As a long-time BeeGFS user, we understand the challenges of managing high-performance storage at scale,” said Prof. Volker Lindenstruth, Goethe University, Frankfurt, Germany. “With our new AMD GPU cluster, we needed exceptional storage performance across diverse workloads, without sacrificing ease of use, reliability, or budget constraints-which all-flash solutions couldn’t meet. VDURA’s hybrid architecture delivered the ideal balance of price/performance and low management overhead, making it the perfect choice.“

Excellence in Partner Collaboration
“Penguin Solutions is excited about VDURA’s new high-performance storage solution. With its microservices-based portability, this new platform aligns well to meet the increased performance, and scalability demands of today’s AI and HPC workloads,” said Dara Ambrose, VP of engineering at Penguin Solutions. “We’re confident when deploying with VDURA’s VeLO and the V5000 that it will meet even the most demanding data challenges, empowering users to push the limits of innovation.”

Industry Analyst
“While modern HPC and AI workloads are driving significant performance requirements, much more is demanded from storage solutions to keep computing resources, both CPUs and GPUs, fully utilized. Reliable storage that can provide data when it’s needed with complete confidence in its integrity is just as important, if not more so, than absolute performance,” observes Mark Nossokoff, research director and storage analyst at Hyperion Research. “VDURA’s focus on coupling enterprise storage solution durability and availability with GPC and AI performance capabilities seeks to address the increasingly demanding requirements of these modern workloads.“