32 x 400Gb QSFP-DD Data Center Switch, 12.8Tbps, Sub-500ns Latency
$12,559.00 VAT excl.
- 12.8 Tbps in 1U: 32×400G QSFP-DD ports for super-spine and spine layers.
- Ultra-low latency: Sub-500 ns port-to-port.
- Advanced telemetry: INT for real-time packet visibility and congestion control.
- High availability: Hot-swappable redundant power and fan modules.
- Open AsterNOS: Enterprise SONiC pre-installed with top-tier SAI support.
- RoCE & AI fabrics: RoCEv2, BGP-EVPN, VXLAN, MLAG, PFC, ECN, DCBX, and multicast collectives.
- Fully compatible: Works with all major GPUs and SmartNICs.
500 ns
End to End Latency
12.8 Tbps
Switching Capacity
70 MB
Packet Buffer
7600 Mpps
Packet Forwarding Rate
Designed for AI inference and training, high-performance computing (HPC), and cloud-scale storage, the Asteraix CX732Q-N delivers 12.8 Tbps switching capacity in a compact 1U form factor with 32×400G QSFP-DD. Built on Marvell Teralynx 7, this switch supports ultra-low latency (Sub-500 ns) and line-rate L2/L3 forwarding. Enterprise SONiC (AsterNOS) is pre-installed for RoCEv2, EVPN multi-homing, and cloud-native operation.
Networking
- Support for all leading collective communication libraries, including NCCL, RCCL, oneCCL, BCL, MPI, and Gloo
- Optimize collective operations (Allreduce, Allgather, All-to-all, Broadcast) with built-in multicast
- Support various network topologies, including spine-leaf, fat-tree, dragonfly, butterfly, 3D mesh/torus, hyper-x, and hypercube
- Full RoCEv2 data path with PFC, ECN, DCBX, and INT for lossless AI fabrics
- BGP-EVPN, VXLAN, and MLAG for scalable leaf-spine and multi-tenant deployments
Use case scenarios
- AI / ML training and inference clusters
- High-performance computing (HPC) fabrics
- Cloud and colocation spine-leaf networks
- High-density data center top-of-rack (ToR)
- Storage networks with RoCEv2 optimization
| Data Plane Ports (400G) | 32x400GbE QSFP-DD | Data Plane Ports (10G) | 2x10GbE SFP+ |
|---|---|---|---|
| Switch ASIC | Marvell Teralynx | Switching Capacity | 12.8Tbps |
| Forwarding Rate | 7600Mpps | Packet Buffer | 70MB |
| Control CPU | Broadwell DE D-1508 | Control Plane Memory | 16GB DDR4, extensible to 96GB |
| ARP Entries | 36K | Mac Addr. Entries | 44K |
| IPv4 Prefix Routes | 72K | IPv6 Prefix Routes | 36K |
| Power Module | 1+1 Hot-swappable | FAN Modules | 5+1 Hot-swappable |
| Input Range | 200- 240V AC / 200-320V DC | Max Power Consumption | 970Watt |
| Rack Space | 1U | Management Ports | 1xUSB2.0, 1xConsole RJ45, 1xMGMT RJ45 |
| Airflow | Front to Rear | Dimension (HxWxD) | 44x440x515mm |
| Op.Temperature | 0 - 40℃ | Op. Humidity | 5% - 90% (non-condensing) |
Loading…
Loading…
Loading…
Loading…
Loading…
| Interface | QSFP-DD |
|---|---|
| Role | Leaf-Spine |
| Speed | 400G |
500 ns
12.8 Tbps
70 MB
↓ 5% vs. IB
Completion Time
↑ 27.5% vs. IB
Token Generation Rate
Advanced Hardware Features
Marvell Teralynx 7 Inside: Unleash Industry-Leading Performance & Ultra-Low Latency
With Marvell Teralynx 7 at its core, 400G switches achieve blazing-fast sub-500ns latency—perfect for AI/ML, HPC, and NVMe where speed drives results.
Redundant Power Supplies & Smart Fans
Boasting 32×400GbE QSFP-DD and 2×10GbE ports, the device supports interface splitting via breakout cables, which helps enhance the efficiency of network resource usage and accommodate diverse application scenarios.
The dual 1+1 hot-pluggable power supplies ensure continuous operation by allowing replacement without system shutdown. Equipped with 5+1 hot-pluggable fan modules, the system delivers reliable cooling performance to maintain stability and efficiency under demanding workloads.
Full RoCEv2 Support for Ultra-Low Latency & Lossless Networking
Asteraix 400G switches deliver full RoCEv2 support with ultra-low latency and near-zero CPU overhead. Combined with PFC and ECN for advanced congestion control, it enables a true lossless network.
Optimized Features for AI/ML/HPC
Enterprise SONiC is optimized for AI/ML/HPC scenarios, supporting efficient and resilient path selection via WCMP (Weighted Cost Multi-Path), as well as technologies like Flowlet to enhance traffic management.
HIGHLIGHT
Network Monitoring & Visualization
AsterNOS supports Node Exporter to send CPU, traffic, packet loss, latency, and RoCE congestion metrics to Prometheus. Paired with Grafana, it enables real-time, visual insight into network performance.
HIGHLIGHT
O&M in Minutes, Not Days
Automate with Easy RoCE, ZTP, Python and Ansible, SPAN / ERSPAN Monitoring & more—cutting config time, errors, and costs.
400G Switch Outperforms InfiniBand in AI Inference with Higher TGR and Lower Latency
In AI inference networks, the 400G switch delivers higher TGR (Token Generation Rate) and lower P90ITL (90th Percentile Inter-Token Latency) compared to InfiniBand, demonstrating faster inference speed and greater overall throughput performance.
