Fully shared packet buffers deliver fair, predictable, and high-bandwidth data paths

Robust RoCE transport powers NVMe over fabric and machine learning applications leveraging NVIDIA GPUDirect

Best-in-class VXLAN scalability with 10x more tunnels and tunnel endpoints than other devices

512K shared forwarding entries, flexibly shared across ACLs, LPM routes, host routes, MAC, ECMP, and tunnel applications

Supports RoCE, Layer-2, and Layer-3

Advanced network virtualization with high-performance single-pass VXLAN routing and IPv6 segment routing

Cloud-scale NAT with 100K+ sessions

Programmable parsing, processing, and editing of packets

Deep packet inspection with 512B depth

Supports WJH, significantly reducing mean time to resolve issues: When, What, Who, Where, and Why

Supports Inband Network Telemetry (INT) and streaming telemetry.

Build a predictable data center with a predictable, affordable network

Zero packet loss

Supports speeds of 1/10/25/40/50/56/100GbE

Easy to deploy

Easy to maintain

Delivers line-rate performance for packets of any size across all ports

Faster storage and server application performance

Supports Software-Defined Networking (SDN)

6.4Tb/s throughput, 47.6 billion packets per second (Bpps)

Up to 32 × 40/56/100GbE ports and 64 × 10/25/50GbE ports in 1RU

300 nanoseconds port-to-port latency for 100GbE

Zero-jitter latency across Layer 2 and Layer 3 forwarding

Build a predictable data center with a predictable, affordable network

Build a predictable data center with a predictable, affordable network — zero packet loss

Supports speeds of 1/10/25/40/50/56/100GbE

Easy to deploy

Easy to maintain

Delivers line-rate performance for packets of any size across all ports

Faster storage and server application performance

Supports Software-Defined Networking (SDN)

4Tb/s throughput, 29.7 billion packets per second (Bpps)

Up to 8 × 40/56/100GbE ports, 48 × 10/25GbE ports, and 16 × 50GbE ports in 1RU

300 nanoseconds port-to-port latency for 100GbE

Zero-jitter latency across Layer 2 and Layer 3 forwarding

Zero packet loss

Easily scales from a single node and switch to thousands of nodes and switches

Supports speeds of 1/10/25/40/50/56/100GbE

Easy deployment

Easy maintenance

Delivers line-rate performance for packets of any size across all ports

Accelerates performance for storage and server applications

Supports Software-Defined Networking (SDN)

3.2Tb/s throughput, 2.38 billion packets per second (Bpps)

Up to 16 x 40/56/100GbE ports, 64 x 10/25GbE ports, and 32 x 50GbE ports in 1RU

300 nanoseconds port-to-port latency for 100GbE

Zero-jitter latency for Layer 2 and Layer 3 forwarding

Less than 6 watts per port

RoHS compliant

Ethernet Storage Fabric (ESF)

Supports RoCE, full buffer utilization, and low latency

Zero packet loss

Supports speeds of 1/10/25/40/50/56/100GbE

Easy to deploy

Easy to maintain

Delivers line-rate performance for packets of any size across all ports

Faster performance for storage and server applications

Supports Software-Defined Networking (SDN)

Throughput: 1.7Tb/s, 12.6 Billion Packets Per Second (Bpps)

100GbE port-to-port latency: 300 nanoseconds

Zero jitter latency for Layer 2 and Layer 3 forwarding

RoHS compliant

Fully shared packet buffers deliver fair, predictable, and high-bandwidth data paths

Consistently low cut-through latency

Robust RoCE transport powers NVMe over fabric and machine learning applications leveraging NVIDIA GPUDirect

Best-in-class VXLAN scalability with 10x more tunnels and tunnel endpoints than other devices

512K shared forwarding entries, flexibly shared across ACLs, LPM routes, host routes, MAC, ECMP, and tunnel applications

Supports RoCE, Layer-2, and Layer-3

Advanced network virtualization with high-performance single-pass VXLAN routing and IPv6 segment routing

Cloud-scale NAT with 100K+ sessions

Programmable parsing, processing, and editing of packets

Deep packet inspection with 512B depth

Deep packet inspection with 512B depth

Supports WJH, significantly reducing mean time to resolve issues: When, What, Who, Where, and Why

Fully shared packet buffers deliver fair, predictable, and high-bandwidth data paths

Consistently low cut-through latency

Robust RoCE transport powers NVMe over fabric and machine learning applications leveraging NVIDIA GPUDirect

Best-in-class VXLAN scalability with 10x more tunnels and tunnel endpoints than other devices

512K shared forwarding entries, flexibly shared across ACLs, LPM routes, host routes, MAC, ECMP, and tunnel applications

Supports RoCE, Layer-2, and Layer-3

Advanced network virtualization with high-performance single-pass VXLAN routing and IPv6 segment routing

Cloud-scale NAT with 100K+ sessions

Programmable parsing, processing, and editing of packets

Deep packet inspection with 512B depth

Supports WJH, significantly reducing mean time to resolve issues: When, What, Who, Where, and Why

Supports Inband Network Telemetry (INT) and streaming telemetry.

Fully shared packet buffers deliver fair, predictable, and high-bandwidth data paths

Consistent and low cut-through latency

Robust RoCE transport powers NVMe over fabric and machine learning applications leveraging NVIDIA GPUDirect

Best-in-class VXLAN scalability with 10x more tunnels and tunnel endpoints than other devices

512K共享转发表项，跨ACL、LPM路由、主机路由、MAC、ECMP和隧道应用灵活共享

Supports RoCE, Layer-2, and Layer-3

Advanced network virtualization with high-performance single-pass VXLAN routing and IPv6 segment routing

Cloud-scale NAT with 100K+ sessions

Programmable parsing, processing, and editing of packets

Deep packet inspection with 512B depth

Supports WJH, significantly reducing mean time to resolve issues: When, What, Who, Where, and Why

支持带内网络遥测 (INT) 、流式遥测

1+1 hot-swappable power supplies and N+1 hot-swappable fans

Color-coded PSUs and fans

Up to 128x 100/50/25/10/1GbE ports, 64x 200GbE ports, or 32x 400GbE ports

Multi-chassis Link Aggregation (MLAG) for active-active Layer 2 multi-pathing

128-way ECMP routing for load balancing and redundancy

Hardware-accelerated histograms track and summarize queue depth with sub-microsecond granularity

In-band Network Telemetry (INT) and streaming telemetry

512K on-chip traffic counters

Fully shared packet buffers deliver fair, predictable, and high-bandwidth data paths

Low latency

Robust RoCE transport powers NVMe by leveraging the switching fabric of NVIDIA GPUs and machine learning applications

Best-in-class VXLAN scalability with 10x more tunnels and tunnel endpoints than other devices

512K shared forwarding entries, flexibly shared across ACLs, LPM routes, host routes, MAC, ECMP, and tunnel applications

Supports Layer-2, Layer-3, and RoCE

Advanced network virtualization with high-performance single-pass VXLAN routing and IPv6 segment routing

Cloud-scale NAT with 100K+ sessions

Programmable pipeline enabling programmable parsing, processing, and editing of packets

Deep packet inspection with 512B depth

Fully shared packet buffers deliver fair, predictable, and high-bandwidth data paths

Consistent and low cut-through latency

Robust RoCE transport powers NVMe over fabric and machine learning applications leveraging NVIDIA GPUDirect

Best-in-class VXLAN scalable tunnels and tunnel endpoints, 10x more than other devices

512K shared forwarding entries, flexibly shared across ACLs, LPM routes, host routes, MAC, ECMP, and tunnel applications

Supports RoCE, Layer-2, and Layer-3

Advanced network virtualization with high-performance single-pass VXLAN routing and IPv6 segment routing

Cloud-scale NAT with 100K+ sessions

Programmable packet parsing, processing, and editing

Deep packet inspection with 512B depth

支持带内网络遥测 (INT) 、流式遥测

Up to 400GbE per port, delivering 25.6Tb/s of aggregate switch bandwidth

Fully shared packet buffers for fair, predictable, and high-bandwidth data paths

Consistent, low cut-through latency

Converged Ethernet acceleration for RDMA (RoCE) and scaling of AI cloud servers built on Ethernet

Best-in-class VXLAN scale

512,000 shared forwarding entries, flexibly shared across ACLs, Longest Prefix Match (LPM) routes, host routes, MAC, ECMP, and tunnel applications

1+1 hot-swappable power supplies and N+1 hot-swappable fans

Ethernet VPN (EVPN) multi-homing

Multi-chassis Link Aggregation (LAG) for active/active L2 multipathing

Hardware-assisted In-Service Software Upgrade (ISSU)

Up to 800GbE per port, delivering 51.2Tb/s of aggregate switch bandwidth

Fully shared packet buffers for fair, predictable, and high-bandwidth data paths

Consistent, low cut-through latency

Converged Ethernet acceleration for RDMA (RoCE) and scaling of AI cloud servers built on Ethernet

Best-in-class VXLAN scale

512,000 shared forwarding entries, flexibly shared across ACLs, Longest Prefix Match (LPM) routes, host routes, MAC, ECMP, and tunnel applications

1+1 hot-swappable power supplies and N+1 hot-swappable fans

Ethernet VPN (EVPN) multi-homing

Multi-chassis Link Aggregation (LAG) for active/active L2 multipathing

Hardware-assisted In-Service Software Upgrade (ISSU)

256-way Equal-Cost Multi-Path (ECMP) routing for load balancing and redundancy