MPN: MSN4700-WS2RC·By: NVIDIA
Ships from the US.
We'll confirm availability and pricing for MSN4700-WS2RC, usually within one business day.
We'll confirm availability and pricing for MSN4700-WS2RC, usually within one business day.
The NVIDIA Mellanox MSN4700-WS2RC is an enterprise-class, Spectrum-3 ASIC-based 400GbE open Ethernet switch designed for high-performance data center deployments. This 1U form factor switch delivers 32 QSFPDD ports supporting flexible port speeds from 1GbE to 400GbE, enabling seamless connectivity across heterogeneous network endpoints. With a switching capacity of 25.6Tb/s and forwarding rate of 8.4Bpps, the MSN4700-WS2RC provides non-blocking, wire-speed performance for demanding leaf/spine/super-spine architectures. Pre-installed with Cumulus Linux, the switch supports comprehensive Layer 2/3 protocols including MLAG, BGP-EVPN, VXLAN, MPLS, and ECMP, along with advanced features like RoCE, PFC, ECN, and deep packet inspection (512B depth). Dual hot-swappable AC power supplies and six N+1 redundant fans ensure high availability. The switch includes 16GB system memory, 64GB SSD storage, and a quad-core x86 CPU for robust management and telemetry. Ideal for cloud-scale deployments, HPC clusters, and AI/ML infrastructure requiring ultra-low latency and massive bandwidth.
| Manufacturer part # | MSN4700-WS2RC |
|---|---|
| Brand | NVIDIA |
| Condition | New |
| Fans | 6 N+1 hot-swappable fans |
| Mounting | Rail kit included for standard 19-inch EIA racks |
| Telemetry | Inband Network Telemetry (INT), streaming telemetry, What-Just-Happened (WJH) diagnostics |
| Model Name | MSN4700-WS2RC |
| SSD Storage | 64GB SSD |
| Flow Control | Priority Flow Control (PFC), Explicit Congestion Notification (ECN) |
| Jumbo Frames | 9,216 bytes supported |
| Product Line | Spectrum SN4000 Series |
The MSN4700-WS2RC is a standalone, enterprise-grade data center switch compatible with standard 19-inch racks, dual AC power infrastructure, and industry-standard QSFPDD optics. Pre-installed Cumulus Linux supports comprehensive Layer 2/3 protocols and SDN integration. Designed for integration into modern data center fabrics (leaf/spine/super-spine) and cloud-scale deployments without requiring proprietary management systems.
Designed for leaf/spine/super-spine data center architectures; compatible with standard 19-inch EIA racks with C2P (front-to-back) airflow.
Backbone switching for large-scale data center fabrics requiring ultra-high bandwidth and low-latency interconnection between leaf switches and core infrastructure.
Multi-tenant cloud environments leveraging VXLAN, EVPN, and cloud-scale NAT (100K+ sessions) for flexible network virtualization and tenant isolation.
Scientific computing clusters and supercomputing facilities requiring 25.6Tb/s non-blocking throughput and 650ns cut-through latency for tightly coupled parallel workloads.
GPU-accelerated machine learning clusters utilizing RoCE (RDMA over Converged Ethernet) with PFC and ECN for lossless, low-latency data movement between compute nodes.
Details
Shipping, warranty & returns
No reviews yet — be the first to review this part.
Secure SSL checkout. Visa, Mastercard, Amex and Discover accepted. ACH / wire available for businesses.
New, certified-refreshed and OEM stock with guaranteed compatibility, backed by warranty on every part.
We process quotes, POs, stock documentation and compliance invoicing.
[email protected]| Product Type | Data Center Ethernet Switch |
|---|
| RDMA Support | RoCE (RDMA over Converged Ethernet) with PFC and ECN |
|---|
| Flow Counters | 512K on-chip flow counters |
|---|
| Packet Buffer | 64MB shared packet buffer |
|---|
| Timing Support | PTP (Precision Time Protocol), SyncE (Synchronous Ethernet), BC Clock Mode with nanosecond precision |
|---|
| Forwarding Rate | 8.4Bpps (billion packets per second) |
|---|
| Programmability | OpenFlow support, programmable packet processing, deep packet inspection (512B depth) |
|---|
| Airflow Direction | C2P (front-to-back) |
|---|
| Layer 2 Protocols | STP, RSTP, MSTP, MLAG, VRRP, LLDP |
|---|
| Layer 3 Protocols | BGP-4, BGP-4+, OSPFv2, OSPFv3, RIP, ECMP, BFD, GR (Graceful Restart) |
|---|
| Forwarding Entries | 512K shared forwarding entries (flexibly shared across ACL, LPM routes, host routes, MAC, ECMP, tunnel) |
|---|
| IPv4 Route Entries | Up to 1M IPv4 route entries |
|---|
| Overlay Networking | VXLAN, EVPN, MPLS, cloud-scale NAT (100K+ sessions) |
|---|
| Switching Capacity | 25.6Tb/s non-blocking full-duplex |
|---|
| Redundancy Features | Hot-swappable power supplies, hot-swappable fans, MLAG support, BFD fast failover, graceful restart (GR) |
|---|
| Switch Architecture | Spectrum-3 ASIC-based open Ethernet switch |
|---|
| Use Case Optimization | Leaf/spine/super-spine data center architectures, cloud-scale deployments, HPC clusters, AI/ML infrastructure, financial services networks |
|---|
| Supported Transceiver Types | QSFPDD 400G, 200G, 100G, 50G, 40G, 25G, 10G, 1G optics; DAC (direct-attach copper) and AOC (active optical cable) compatible |
|---|
| System Memory | 16GB RAM |
|---|
| Power Supplies | Dual 1+1 hot-swappable AC power supplies (100-264VAC, 50-60Hz) |
|---|---|
| Maximum Power Consumption | 800W |
| Typical Power Consumption | 630W (ATIS) |
| Rack Units | 1 RU |
|---|---|
| Form Factor | 1U rackmount, standard depth |
| Connector Type | QSFPDD (QSFP Double Density) |
|---|---|
| Maximum 1GbE Ports | 128 (breakout mode) |
| Port Configuration | 32x QSFPDD 400GbE ports |
| Maximum 10GbE Ports | 128 (breakout mode) |
| Maximum 25GbE Ports | 128 (breakout mode) |
| Maximum 40GbE Ports | 64 (breakout mode) |
| Maximum 50GbE Ports | 128 (breakout mode) |
| Management Interface | Out-of-band management via quad-core x86 CPU with 16GB memory and 64GB SSD |
| Maximum 100GbE Ports | 128 (breakout mode) |
| Maximum 200GbE Ports | 64 (breakout mode) |
| Maximum 400GbE Ports | 32 |
| Supported Port Speeds | 400G, 200G, 100G, 50G, 40G, 25G, 10G, 1G per-port flexible |
| Operating System | Cumulus Linux pre-installed |
|---|
| Cut-Through Latency | 650ns consistent |
|---|
| CPU | Quad-core x86 2.20GHz Intel processor |
|---|
| Dimensions | 1.72" (H) x 16.85" (W) x 22.3" (D) / 44mm (H) x 428mm (W) x 568.5mm (D) |
|---|
| Environmental Compliance | RoHS compliant |
|---|
| Operating Humidity | 5% to 95% non-condensing |
|---|---|
| Storage Temperature | -40°C to 70°C |
| Operating Temperature | 0°C to 45°C |
Ultra-low latency trading and financial data networks requiring consistent 650ns latency, PTP/SyncE precision timing, and deterministic packet forwarding.
Next-generation data center upgrades consolidating multiple legacy switch tiers into single unified 400GbE fabric with flexible port speed adaptation.
The MSN4700-WS2RC delivers 25.6Tb/s non-blocking switching capacity with 8.4Bpps (billion packets per second) forwarding rate. This enables wire-speed, full-duplex performance across all 32 ports simultaneously, supporting demanding applications like HPC, AI/ML, and cloud-scale data centers without performance bottlenecks.
Yes. The switch provides comprehensive RoCE (RDMA over Converged Ethernet) support with Priority Flow Control (PFC) and Explicit Congestion Notification (ECN) for lossless, low-latency data movement. This enables efficient GPU-to-GPU communication in AI/ML clusters and HPC environments, leveraging NVIDIA GPUDirect for intelligent hardware-accelerated congestion management.
Each of the 32 QSFPDD ports supports flexible speeds: 400G, 200G, 100G, 50G, 40G, 25G, 10G, and 1G. This allows seamless connectivity to heterogeneous endpoints without requiring separate switch tiers, simplifying network architecture and reducing operational complexity in mixed-speed environments.
The switch runs Cumulus Linux, a Linux-based NOS supporting comprehensive Layer 2/3 protocols: BGP-EVPN, VXLAN, MPLS, ECMP, MLAG, STP/RSTP/MSTP, OSPFv2/v3, and RIP. It also supports OpenFlow for SDN integration and cloud-scale NAT with 100K+ concurrent sessions, enabling flexible network virtualization and programmability.
The MSN4700-WS2RC includes dual 1+1 hot-swappable AC power supplies (100-264VAC, 50-60Hz) and six N+1 hot-swappable fans. It supports MLAG for multi-chassis link aggregation, BFD for fast failover, and graceful restart (GR) for non-disruptive maintenance, ensuring carrier-grade availability for mission-critical deployments.
The MSN4700-WS2RC achieves 650ns cut-through latency, among the lowest in the industry. This consistent, predictable latency is critical for ultra-low-latency applications like financial trading, real-time analytics, and tightly coupled HPC workloads where microsecond-level delays impact performance and competitiveness.
The switch includes Inband Network Telemetry (INT)-ready hardware, streaming telemetry, 512K on-chip flow counters, and What-Just-Happened (WJH) diagnostics. Deep packet inspection up to 512 bytes enables comprehensive traffic analysis, reducing mean-time-to-resolution (MTTR) for network issues and supporting advanced monitoring and analytics.
You might also want to check out these products.

Cisco Meraki MS125-48LP-HW

Cisco Meraki MS125-48LP

AT-8624T/2M - Allied Telesis 8600 24-Port 2x Module Bays uplink Switch

AT-9000/24 - Allied Telesis 9000/52-10 24-Port Switch

469-3414 - Dell PowerConnect 5524 24-Port + 2 x SFP+ Ports 1U Switch

AT-GS910/24-50 - Allied Telesis CentreCOM GS910 24-Ports Switch

AT-FS717FC/SC-10 - Allied Telesis 16x 1x 100BFX SC Connector Switch

AT-9424T/SP - Allied Telesis 9400 24-Port 2x Port SFP Switch