HighPoint Technologies, a pioneer in high-performance NVMe storage infrastructure, and OSNexus, a leading developer of unified Software-Defined Storage (SDS) solutions, have announced the official qualification of OSNexus QuantaStor on the HighPoint RocketStor 4243AS (RS4243AS) NVMe-over-Fabrics (NVMe-oF) storage platform.
This synergistic turnkey solution redefines Composable Disaggregated Infrastructure (CDI) by combining an open, Bring-Your-Own-Drive (BYOD) storage hardware platform with an enterprise-class software management grid. The result is a True Unified Storage pipeline that simultaneously serves file (NFS/SMB), block (iSCSI/NVMe-oF), and object (Amazon S3-compatible) workloads over standard 100GbE Ethernet infrastructure—completely eliminating the high costs and rigid vendor lock-in associated with proprietary flash arrays.
The Power of Synergy: Bare-Metal Performance Meets Grid Management
Modern data-intensive industries are frequently trapped between two extremes complex, expensive legacy storage infrastructure, or fragmented software tools that are difficult to manage. HighPoint and OSNexus’ collaboration effectively resolves this dilemma by cleanly separating high-density NVMe hardware fabric acceleration from the unified software control layer.
The HighPoint RocketStor 4243AS packs up to 24 enterprise-class U.2/U.3 NVMe SSDs into a compact industry-standard 2U rackmount footprint. It offloads the entire NVMe-oF protocol stack to dedicated fabric bridge silicon, routing clean, low-latency storage traffic over dual 100GbE ports using NVMe/TCP or ultra-low-latency RoCE v2 fabrics.
OSNexus QuantaStor acts as the master intelligence engine over this high-speed pool. It ingests the raw NVMe-oF block capacity and natively translates it into diverse multi-protocol storage tiers. Through QuantaStor’s advanced Storage Grid Technology, administrators can manage these high-density appliances alongside multi-site data centers and hybrid public clouds from a single, federated management interface.
Key Value Propositions for IT Architects
- End-to-End CPU Offload via RoCE v2 Lowers TCO: When deployed over a RoCE v2 network fabric, the solution achieves total, dual-sided CPU offload to deliver massive operational cost savings. The RocketStor 4243AS processes the NVMe-oF protocol entirely in hardware silicon at the target side, while hardware-driven RDMA completely bypasses the host server’s network stack on the initiator side. By eliminating this storage-related CPU overhead, CSPs and MSPs can instantly reclaim valuable host compute cycles to maximize virtual machine (VM) density and lower data center TCO.
- Unified Enterprise Storage Pool: Run high-throughput file shares, sub-millisecond database block storage, and massive S3-compatible object repositories simultaneously from a single unified pool of enterprise NVMe storage.
- Maximizing Capital Efficiency with a BYOD Architecture: Engineered as a hardware-neutral platform, the solution optimizes storage performance and density-per-dollar by allowing organizations to source industry standard commercially available SSDs at market prices. The streamlined, space-efficient 2U rackmount chassis has been engineered to maximize raw NVMe-oF throughput without the cost or complexity of conventional CDI infrastructure.
- Single-Pane-of-Glass Enclosure Management: The joint solution delivers deep hardware-to-software visibility. Administrators can directly manage and monitor fan speeds, power supplies, and view a visual, direct slot-to-drive mapping of the physical RS4243AS chassis through the QuantaStor management console.
- Protocol Flexibility (NVMe/TCP and RoCE v2): Organizations can customize deployments using dual 100GbE connectivity. It supports low-cost NVMe/TCP setups over standard networking fabrics or ultra-low-latency, lossless RoCE v2 (RDMA over Converged Ethernet) pipelines.
Target Applications & Industry Impact
AI, Machine Learning, and Data Science Pools
High-performance GPU clusters often sit idle while waiting for traditional storage systems to feed them training datasets. This solution effectively eliminates “data stalling” by delivering millions of IOPS over 100GbE fabrics; keeping expensive GPU nodes continuously saturated at local-flash speeds. QuantaStor presents this data as a massively scalable S3-compatible object storage lake or an NFS mount, enabling seamless, high-speed ingestion and checkpointing.
Media & Entertainment (M&E) Post-Production
Editing multi-stream 8K RAW video requires massive sequential throughput across collaborative teams without dropping frames. Production houses can deploy the joint solution as a high-velocity central NAS via standard SMB/NFS. Creative teams can edit high-bitrate video directly over the network fabric, while the hardware-neutral design allows studios to rapidly scale up raw capacity on demand as project sizes expand.
Private Cloud Virtualization and Hypervisors
By adopting a disaggregated CDI approach, IT departments can scale high-density RocketStor NVMe storage nodes completely independently of their compute layers. Hypervisors pull high-performance block storage targets over standard Ethernet fabrics, maximizing infrastructure flexibility and hardware utilization while eliminating stranded hardware resources.
Executive Perspectives
“HighPoint’s RocketStor 4243AS provides a great building block for our scale-out Ceph based storage clusters and enables organizations to expand any server from any manufacturer with additional NVMe storage capacity at any time,” said Steven Umbehocker, Founder & CEO of OSNexus. “With current supply chain challenges around NAND/flash media this is coming at a key time when flexibility to expand existing infrastructure is paramount.”
“The RocketStor 4243AS was purposefully engineered to democratize NVMe-over-Fabrics by utilizing a streamlined, single-controller architecture that maximizes thermal efficiency, power savings, and raw storage density per rack unit,” said May Hwang, VP of Marketing at HighPoint Technologies, “Partnering with OSNexus allows us to wrap that ultra-low-latency hardware fabric in a mature, enterprise-grade data management layer. Our customers in AI, media production, and private cloud data centers can now leverage a true unified storage solution that handles file, block, and object data seamlessly at full 100GbE line rates without a proprietary cost premium.”
