Many businesses need fast AI processing but their current servers are slow and take up too much space. When the hardware cannot keep up, projects stall and costs rise. The Firefly CSB2-N10NOrinNX, 1570 TOPS, 16GB LPDDR5, 2U Rackmount Server gives you a compact, powerful solution that fits standard racks and speeds up AI workloads.
Firefly CSB2-N10NOrinNX, 1570 TOPS, 16GB LPDDR5, 2U Rackmount Server with Firefly
Meet the server that turns heavy AI jobs into smooth operations. It is a 2U rack-mount machine built on ARM architecture with ten distributed compute nodes. Each node runs an NVIDIA Jetson Orin NX processor, delivering a total of 1570 TOPS of AI power. If you are struggling with laggy video analytics or slow model training, this Firefly system removes the bottleneck and lets you work faster.
Key Features
- 10 compute nodes plus one control node on a single 2U chassis.
- Octa-core 64-bit NVIDIA Jetson Orin NX processor at up to 2.0 GHz.
- Total AI compute power of 1570 TOPS (INT8) across all nodes.
- ARM Cortex-A78AE CPU for efficient general-purpose tasks.
- NVIDIA Ampere GPU with 1024 CUDA cores and 32 Tensor Cores per node.
- Supports up to 196 AI video channels for complex surveillance.
- 256 GB PCIe NVMe SSD plus six hot-swap SATA slots for flexible storage.
- 1 Half-height PCIe 2.0 4 slot for additional expansion cards.
- Four high-speed cooling fans ensure stable operation in 0 C-42 C environments.
- Optional redundant dual power supplies for maximum uptime.
- Two 10 Gb Ethernet SFP+ ports, two Gigabit RJ45 ports, and a dedicated management port.
- Standard 19-inch rack mounting with IEC 297 universal cabinet compatibility.
Specifications
| Product Model |
CSB2-N10NOrinNX |
| Server Form |
2U rack-mounted computing power server |
| Architecture |
ARM architecture |
| Number of Nodes |
10 distributed computing nodes + 1 control node |
| Compute Nodes |
Octa-core 64-bit NVIDIA Jetson Orin NX processor, main frequency up to 2.0GHz |
| AI Computing Power |
1570 TOPS (157 TOPS × 10 nodes, INT8) |
| CPU |
Octa-core 64-bit ARM Cortex-A78AE v8.2 processor, up to 2.0GHz |
| GPU |
NVIDIA Ampere architecture GPU with 1024 CUDA cores and 32 Tensor Cores (per Orin NX node) |
| Video Encoding |
1 × 4K@60fps, 3 × 4K@30fps, 6 × 1080P@60fps, 12 × 1080P@30fps |
| Video Decoding |
1 × 8K@30fps, 2 × 4K@60fps, 4 × 4K@30fps, 9 × 1080P@60fps, 18 × 1080P@30fps |
| AI Video Processing |
Supports up to 196-channel AI video processing (depending on model and core configuration) |
| RAM |
16GB LPDDR5 × 10 nodes |
| Storage |
256GB PCIe NVMe SSD (2242 format, internally assembled) |
| Storage Expansion |
6 × 3.5-inch / 2.5-inch SATA 3.0 HDD/SSD slots, supports hot swapping |
| PCIe Expansion |
1 × Half-height half-length PCIe 2.0 ×4 standard slot (signal rate: PCIe 2.0 ×1) |
| Power Supply |
Single power supply standard; redundant dual power supplies optional |
| Cooling |
4 × high-speed cooling fans |
| Dimensions |
487.9mm × 530.59mm × 88.8mm |
| Installation Requirement |
IEC 297 universal cabinet installation, 19-inch width, cabinet depth ≥800mm;
slide rail installation distance: 543.5mm–848.5mm
|
| Weight |
Server net weight: 11.5kg; packaged weight: 13.7kg |
| Operating Temperature |
0℃ to 42℃ |
| Storage Temperature |
-40℃ to 60℃ |
| Operating Humidity |
5% to 80% RH (non-condensing) |
| Network Interfaces |
2 × 10Gb Ethernet SFP+ ports
2 × Gigabit Ethernet RJ45 ports
1 × Gigabit Ethernet RJ45 management port (BMC)
|
