NVIDIA MSN2010-CB2F Ethernet Switch for Hyperconverged Infrastructures with Onyx
Earn 0 points when you buy me!
Half‑width 1U open‑Ethernet switch (MSN2010‑CB2F) with 18 × 10/25 GbE SFP28 ports + 4 × 40/100 GbE QSFP28 uplinks — designed for hyperconverged, storage‑fabric, and data‑center deployments.
Mellanox NVIDIA Spectrum SN2010 (MSN2010-CB2F) Resources
High-performance multi-speed switching
The MSN2010-CB2F delivers high-performance 10/25/40/50/100 GbE switching in a compact 1U form factor. Designed for hyperconverged and enterprise data centers, it provides ultra-low latency and non-blocking throughput, enabling consistent, predictable performance across cloud, storage, and AI workloads. Hardware-accelerated telemetry ensures rapid network monitoring and troubleshooting.
Flexible open-networking design
The SN2010 supports NVIDIA ONIE boot loader and multiple NOS options including Cumulus Linux, SONiC, and NVIDIA Onyx. Its standards-compliant, open networking architecture allows seamless integration into existing data center fabrics while providing deterministic, lossless traffic for hyperconverged and high-density environments.
Operational efficiency & reliability
The Spectrum-based architecture optimizes power usage and reduces operational complexity. With redundant PSUs and fan modules, the SN2010 ensures continuous uptime for enterprise fabrics, storage networks, and AI clusters, delivering reliable performance and simplified network management for demanding data center workloads.
| MFG Number | 920-9N110-00F7-0X2 |
|---|---|
| Condition | Item Condition : Brand New |
| Price | $55,250.00 |
| Ports | 18 x 25 Gigabit SFP28 + 4 x 100 Gigabit QSFP28 |
| Switches Rack Unit | 1U |
| Capacity | 1.76 Tbps/1,309 Mpps |
| Software | Cumulus Linux |
| Product Card Description | Mellanox Spectrum MSN2010‑CB2F delivers up to 850 Gb/s switching throughput and ~1.26 billion packets per second forwarding performance in a compact half‑width 1RU form factor. It supports 10/25/40/100 GbE data rates with ultra‑low latency (300 ns cut‑through), a shared 16 MB packet buffer for lossless performance, and open‑Ethernet flexibility via ONIE, supporting Cumulus Linux, NVIDIA Onyx or other NOS. Ideal for hyper‑converged infrastructure, storage fabrics, AI/ML clusters, cloud racks, and enterprise data centers requiring high density and low latency. |
| Order Processing Guidelines | Order Processing Guidelines:
Inquiry First – Please reach out to our team to discuss your requirements before placing an order. |