NVIDIA Spectrum SN4600

NVIDIA Spectrum SN4600

Scalable, High-Density Switching for Modern Data Centers

The NVIDIA Spectrum SN4600 is part of the cutting-edge Spectrum SN4000 series, NVIDIA’s 4th generation of advanced data center switches. Designed for high-performance environments, the SN4600 is purpose-built to serve in spine, super-spine, and high-density leaf roles within cloud-scale, AI/ML, and enterprise-grade data centers.

This 2U switch offers 64 ports of 200GbE (or 128 ports of 10/25/50GbE via splitter cables), providing unmatched port flexibility and high throughput for network architects aiming to deploy highly efficient and scalable infrastructure.

With support for Onyx™, Cumulus Linux®, and ONIE, and advanced SDN features, SN4600 is engineered to meet the requirements of both traditional and modern disaggregated networks.

Key Features at a Glance

• 64x 200GbE ports in a dense 2U form factor
• Up to 4 Tb/s throughput
• Landmark 4 Billion packets per second (Bpps) processing capacity
• Support for split configurations up to 128x 10/25/50GbE
• Compatible with QSFP56 and QSFP28 transceivers
• Flexible operating system support: Onyx, Cumulus Linux, ONIE
• Ideal for spine, super-spine, and high-density leaf deployments
• Low-latency, non-blocking architecture for cloud-native applications
• SDN-ready, with full support for automation, telemetry, and DevOps toolchains
• Advanced layer 2/3 protocols and next-gen virtualization support

Design & Architecture

Compact 2U Form Factor, Massive Capability

Despite its 2U size, the SN4600 provides exceptional port density, scalability, and performance.

The switch supports 64 ports of 200GbE, but with splitter cables, each can be subdivided into 2 ports of 100GbE or 4 ports of 10/25/50GbE, allowing for up to 128 endpoints in a single switch.

This makes it the ideal solution for spine connectivity in high-traffic, scale-out architectures.

Flexible Port Speed Ranges

The SN4600’s versatility in port speeds from 1GbE to 200GbE makes it suitable for a wide variety of deployment scenarios.

Whether connecting legacy servers or enabling next-gen compute and storage nodes, this switch can handle it all with zero compromise on performance or consistency.

Performance

Wire-Speed Switching

The SN4600 delivers a throughput of 6.4Tb/s and a processing capacity of 8.4Bpps, ensuring wire-speed switching across all ports under full traffic loads.

This is essential for maintaining performance in latency-sensitive environments like AI training, real-time analytics, and high-frequency trading.

Ultra-Low Latency and Deterministic Performance

Built on NVIDIA’s Spectrum ASIC, the SN4600 guarantees predictable, low-latency forwarding, even under mixed traffic conditions and congested environments.

Its architecture ensures fairness and headroom, avoiding performance drops due to buffer overflow or head-of-line blocking.

Operating System Support

The SN4600 is designed to support open networking and offers seamless integration with leading data center network operating systems:

• NVIDIA Onyx™: Advanced OS with extensive Layer 2/3 support, automation, and telemetry tools.
• Cumulus Linux®: Industry-standard, Linux-based NOS for DevOps-driven environments.
• ONIE: Open Network Install Environment for flexible OS deployments and fast provisioning.

These options allow network teams to choose the best software stack for their operational models, be it automation-first, traditional CLI, or full-scale DevOps integration.

Scalability & Use Cases

Ideal for Spine & Super-Spine Architectures

With its high port density and 200GbE speeds, the SN4600 is ideal as a spine or super-spine switch in leaf-spine topologies, where east-west traffic between racks or clusters must be handled with low latency and high bandwidth.

Perfect as a High-Density Leaf

The SN4600 is also suitable as a high-density leaf switch, capable of connecting hundreds of compute or storage nodes.

This is especially beneficial for high-performance computing (HPC), AI workloads, and hyperconverged infrastructures.

Cloud, AI, and Data-Intensive Environments

Whether you’re operating a hyperscale data center or running large-scale cloud workloads, the SN4600 delivers the non-blocking architecture, deterministic performance, and software flexibility needed for:

• Public and private cloud infrastructure
• Big data and analytics platforms
• AI/ML training clusters
• Content delivery networks (CDNs)
• Financial and trading applications

Advanced Networking Features

• VXLAN & EVPN Support: Enables scalable layer 2 overlays over layer 3 networks, ideal for multitenant cloud environments.
• Automation & Telemetry: Deep visibility into network traffic, congestion, and performance via streaming telemetry and open APIs.
• OpenFlow & SDN Integration: Seamless integration with SDN controllers for centralized management and automation.
• IPv6 & Multicast Ready: Future-proofed networking with full support for dual-stack IP deployments.
• MPLS, BGP, OSPF, and More: Extensive support for traditional routing and modern overlay protocols.

Product Options

SKU	Model	Data Rate	Network Interface	Ports	Technology
920-9N302-00F7-0C2	MSN4600-CS2FC	100GbE	QSFP28	64x100GbE	Ethernet
920-9N302-00R7-0C0	MSN4600-CS2RC	100GbE	QSFP28	64x100GbE	Ethernet
920-9N302-00FA-0C0	MSN4600-VS2FC	200GbE	QSFP56	64x200GbE	Ethernet
920-9N302-00RA-0C0	MSN4600-VS2RC	200GbE	QSFP56	64x200GbE	Ethernet

Why Choose NVIDIA Spectrum SN4600?

• Performance without compromise – Full line-rate forwarding at up to 200GbE per port.
• Port flexibility – From legacy 10GbE to high-speed 200GbE—all in one switch.
• Future-proof your data center – Open architecture and SDN-ready platform.
• Reduced TCO – Efficient power and cooling design, rack-space optimization.
• Operational efficiency – Supports DevOps workflows, zero-touch provisioning, and scalable automation.
• Security & Visibility – Advanced traffic inspection, role-based access control, and rich telemetry for full network insight.

Conclusion

The NVIDIA Spectrum SN4600 is a powerful and versatile switching solution engineered for the most demanding modern data centers.

With unmatched port density, ultra-low latency, robust software support, and open networking compatibility, it empowers organizations to scale faster, operate more efficiently, and prepare their networks for the next generation of cloud, AI, and data workloads.

Whether deployed as a core spine switch or a high-density leaf, the SN4600 delivers unparalleled speed, scalability, and simplicity backed by NVIDIA’s proven excellence in high-performance networking.