HPC data storage stalwart DDN today announced a software-defined storage platform, Infinia, which the company says leverages two decades of engineering experience in file system, data orchestration and AI-based optimization to support the era of accelerated computing and generative AI.
Aimed at cloud platforms and large HPC, hyperscale and data center environments, Infinia combines multi-tenancy at scale, containerization and high performance. “The new architecture accelerates and simplifies workflows for today’s and tomorrow’s data management requirements, from generative AI and large language models (LLM) to versatile and complex movement of workflows from the edge to data centers and the cloud,” DDN said.
Infinia is initially available for Amazon S3 object storage environments, core features include:
- Automated installation configuration – DDN said the platform takes 10 minutes to deploy with downtime reduced via rolling upgrades and elastic expansion; stock allocation is handled by business policy; self-service with multiple tenants
- Data provisioning is simplified — the data engine creates a new tenant in one command, according to DDN; sets capacity limits and performance shares
- Multi-tenancy and QoS – software-based native SLA manager automates distributed key space securely designed for efficient resource sharing
- DDN said Infinia offers scalable metadata management with scalable storage for data management and control while working to reduce complexity and duplication of data and metadata; tracks model training, tag and search, end-to-end model and control
DDN also launched a device for the Infinia platform, a 1U, all QLC flash system with 12 hot-swap 2.5” U.2 NVMe drive bays and up to 4x 200G network ports. Infinia scales out by adding servers according to workload and performance requirements.
DDN said it will also support a wide range of processors and hardware architectures for software-defined implementation in future releases.
DDN Infinia is also available for deployment on Google Cloud, with additional cloud support coming in the future.
The company said Infinia is designed to help enterprises manage disparate data sources in multiple locations – on-premises, at the edge and in multiple cloud environments – as well as disparate systems for different applications and shared access to data sources. Data is created at the edge and must be moved between the edge and the data center to be acted upon. The increasing demand for distributed unstructured data requires metadata tagging to ensure it moves quickly, efficiently and securely.
“Data management has become terribly complex, and the market has been crying out for a new paradigm that addresses the challenges of explosive AI data growth in a fully secure, multi-tenant, power-efficient and cost-effective manner,” said Dr. James Coomer, SVP of Products, DDN. “Enterprises need to aggregate disparate data types collected by many sensors and sources across distributed locations and ensure that the resulting data is manageable, secure and accessible for business and data processing needs. DDN Infinia was built from the ground up to deliver on these needs.”