# VDURA adds multi-tenancy and faster S3 to PanFS > Source: > Published: 2026-06-19 15:27:07+00:00 OBJECT # VDURA adds multi-tenancy and faster S3 to PanFS Parallel file system supplier [VDURA](https://www.blocksandfiles.com/flash/2026/04/24/vdura-ceo-sees-the-end-of-the-all-flash-array-era/5218827) says it has made its software more enterprise and AI-capable by adding a control plane with multi-tenancy, speeding up S3 object storage and adding native S3 object tagging. VDURA says its next-generation control plane is purpose-built for the operational demands of modern AI and HPC environments, with a modern management interface providing a streamlined tenant administration model. It has a new user interface and a REST API for platform operations giving operators and integrators a foundation for automation and tooling workflows. It is also delivering S3 performance improvements designed to sustain high throughput across cloud-native AI pipelines, including AI model checkpointing, inference serving, and large-scale dataset ingestion. These improvements reduce latency for S3-native operations and increase aggregate throughput for concurrent read and write workloads. Chris Girard, vice president of product management at VDURA, said: "The new management interface brings clarity and control to the people running VDURA day-to-day, while our expanded S3 capabilities give data engineers and AI practitioners the tools to build more sophisticated, automated data pipelines. These are the operational and integration capabilities our customers have been asking for." How is S3 being accelerated? Eric Salo, VDURA's SVP of marketing and business operations, said: "We're ultimately a scale-out cloud kind of architecture. So we've got stateless directors that are the brains of the system, and then we've got these storage nodes. One of our first decisions at VDURA was to virtualize the software. So we ran these microservices. We virtualized the S3 servers and now we can put more of these directors in and they all run as agents. And so, if you want to double your S3 capacity, now you just double the number of directors. The directors are cheap compared to the flash and we can scale the S3 performance to almost anything you want." Is VDURA using [S3 over RDMA](https://www.blocksandfiles.com/object/2022/10/21/s3-over-rdma/1616260?_gl=1*1azvxgc*_ga*MzE5OTg0NTIxLjE3NjI4MTAzNjQ.*_ga_NSDTXHMMN0*czE3ODE4NDk5NTEkbzQzOCRnMSR0MTc4MTg1MDAxOCRqNTgkbDAkaDA.)? Salo said: "We are not using S3 over RDMA. We do have RDMA from the clients to the storage nodes. Each one of our storage nodes, we're getting between 60 and 90 gigabytes per second of transfer rate. RDMA really helps us in that." VDURA's PanFS software "is parallel, fully parallel from the beginning, it's a foundational thing for us." So if you have a system with a need for terabytes-per-second bandwidth levels "what we do is we just put these S3 agents in the directors and we put enough of them in there to give you the performance that you want." "We can scale up to pretty high numbers. And what we're finding is, if I just look at how people are actually using this, they don't want direct client levels of S3, they just want to be able to move their… data off at a reasonable clip. And so this has been a really good solution for them. In fact, we actually developed this in conjunction with a couple of the neoclouds we're working with." VDURA is working with AI-focused neoclouds? Yes, it is. Salo wasn't able to publicly name then, but did say: "We've had a couple of really good ones and it's helping us because as we are working with these guys, engineer to engineer, we're learning really quickly about what they need and how their systems really need us to look." VDURA is also introducing native S3 object tagging support, so customers can attach rich metadata to objects stored in VDURA systems. These S3 tags enable policy-based data lifecycle management, automated tiering workflows, and fine-grained access controls across AI training datasets, model artifacts, and scientific data collections. They provide the metadata infrastructure required to enforce governance policies and manage data throughout their full lifecycles for HPC and AI operators managing petabyte-scale object stores. The multi-tenant control plane, S3 performance improvements, and S3 tagging support are planned for general availability for all V5000 class systems in the second half of this year. Existing customers can upgrade in place via an online software update. ®