Parallel file system provider VDURA has upgraded its software with enterprise and AI-focused enhancements, including a multi-tenant control plane, accelerated S3 object storage performance, and native S3 object tagging capabilities.
Its next-gen control plane is tailored for modern AI and HPC operational requirements, featuring a streamlined multi-tenant admin model, a revamped UI, and platform-operation REST API. The updates equip operators and integrators with robust foundations for automation and custom tooling workflows. Optimized for cloud-native AI pipelines including model checkpointing, inference serving and large-scale dataset ingestion, the enhanced S3 storage cuts operation latency and boosts aggregate throughput for concurrent read-write workloads.
“The new interface simplifies daily operation and strengthens control, while upgraded S3 tools empower teams to build advanced, automated data pipelines — addressing core customer demands for better operability and integration,” said Chris Girard, VDURA VP of Product Management.
VDURA’s S3 performance boost stems from its scale-out cloud architecture, explained Eric Salo, SVP of Marketing and Business Operations. The virtualized microservices-based system leverages stateless director nodes as its core brain paired with modular storage nodes. Virtualized S3 server agents run on directors, enabling linear performance scaling: users can double S3 capacity and throughput simply by adding more cost-effective directors, delivering nearly unlimited scalability.
Salo confirmed the platform utilizes client-to-storage-node RDMA (not S3-over-RDMA), enabling each storage node to achieve 60–90 GB/s transfer speeds. Built on fully parallel PanFS software, the system deploys sufficient S3 agent directors to deliver TB/s-level bandwidth for high-demand workloads. The solution was co-developed with several unnamed AI-focused neocloud partners, helping VDURA rapidly iterate to meet real-world cloud AI infrastructure needs.
The new native S3 object tagging function allows users to attach rich metadata to stored objects. It supports policy-driven data lifecycle management, automated tiering, and fine-grained access control for AI training data, model artifacts and scientific datasets, delivering essential metadata infrastructure for governance and full-lifecycle management of petabyte-scale object storage in HPC and AI environments.
The multi-tenant control plane, S3 performance upgrades and object tagging features will be generally available for all V5000-class systems in H2 2025. Existing users can adopt the updates via seamless online in-place upgrades.
Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!
Votre message doit contenir entre 20 et 3 000 caractères!
French
