The Supermicro EBox is powered by H14 CloudDC servers, a purpose-built, all-flash storage platform optimized for VAST Data’s Disaggregated Shared-Everything (DASE) architecture. Combined with VAST AI OS which unifies the datastore, database and AI engine into a single platform, the EBox forms a storage backbone for data-intensive workloads and hyperscale multi-tenant storage.
Each EBox runs both CNode (compute) and DNode (storage) containers on the same server, eliminating the need for separate storage enclosures while preserving DASE’s ability to scale compute and capacity independently across the NVMe fabric. This architecture delivers linear scale-out from terabytes to exabytes under a single global namespace, with inline encryption, deduplication, and erasure coding handled natively with no external accelerators required. Security governance follows data end-to-end: access controls on source files propagate automatically through embeddings to inference outputs. High-bandwidth RDMA networking (200 GbE via NVIDIA® ConnectX®-7/8) enables zero-copy data movement between EBox nodes and clients. For organizations adopting enterprise AI, the same EBox infrastructure supports the VAST InsightEngine for automated RAG pipelines and the native vector database for semantic queries.
The CNode-X AI Data Platform solution with Supermicro, NVIDIA and VAST Data provides a turnkey AI storage and workflow solution for enterprise AI deployment. This solution integrates All-Flash storage arrays, high performance servers, GPU accelerators with VAST’s AI OS which provides universal storage and integrated vector database, event streaming services, and low-code data pipeline builder for orchestrating AI workloads.
The CNode-X solution follows NVIDIA's AI Data Platform reference architecture, which incorporates GPU acceleration into the storage platform to accelerate data vectorization, vector database searching, and inference workloads. The benefits of this solution include:
Total AI Software Stack
Leverage the integrated hardware powered by NVIDIA and VAST provide the comprehensive software stack:
VAST Accelerated Core
- Scalable, Event-Driven GPU Compute for AI pipelines
- Real-Time RAG and Multimodal workflows
- VAST Container Orchestration (RKE2) for NVIDIA NIM
VAST Compute Core
- Transactional, Analytical & Vector Database Services
- VAST Event Broker For Distribution Event Streaming
VAST Capacity
- Globally Distributed DataStore & KV Cache Services
- Universal Storage (NFS, S3, SMB, NVMe over TCP)
CloudDC A+ Server
1U Single-processor CloudDC with DC-MHS, up to 12 hot-swap 2.5" drives, and 3 PCIe 5.0 slots
- Single AMD EPYC™ 9005/9004 Series Processor
- 12 DIMMs (1 DPC); Up to 3TB ECC DDR5 RDIMM with EPYC 9005 series processor
- Flexible networking options with 1 AIOM networking slot (OCP NIC 3.0 compatible)
- Expansion slots: 2 PCIe 5.0 x16 FHHL; 1 PCIe 5.0 AIOM networking slot
- 1000/2000W Redundant Titanium-Level Power Supplies (*Full redundancy based on configuration and application load)
AS -1116CS-TN
GPU SuperServer
2U Single-processor Server with up to 4 NVIDIA® RTX PRO™ 6000 GPUs
- Single Intel® Xeon® 6700/6500 Series Processor with P Cores
- Support for up to 4 double-width PCIe GPU accelerator cards
- Up to 16 DIMMs supporting up to 1TB DDR5-6400 in 1DPC or 2TB DDR5-5200 in 2DPC or 512GB DDR5-8000 in 1DPC
- Up to 4 PCIe 5.0 x16 FHFL double-width + 3 PCIe 5.0 x16 FHFL slots
- Up to 4 rear hot-swap E1.S NVMe drive bays
- 2700W Redundant Titanium-Level Power Supplies
