/
32 x 100Gb QSFP28 Data Center Switch, 3.2Tbps, Sub-500ns Latency
$7,249.00
- 3.2 Tbps in 1U: 32×100G QSFP28 with 40G QSFP+ breakout support.
- 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
3.2 Tbps
Switching Capacity
70 MB
Packet Buffer
6300 Mpps
Packet Forwarding Rate
Designed for AI inference and training, high-performance computing (HPC), and cloud-scale storage, the Asteraix CX532P-N delivers 3.2 Tbps switching capacity in a compact 1U form factor with 32×100G QSFP28. 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
SONiC-VPP Routing OS
AsterNOS-VPP is a production-hardened SONiC routing OS with VPP as the data-plane engine. The full feature set ships pre-activated on RT2500 series routers — no per-feature license keys, no add-on modules to purchase separately.
Interface
Routing Protocols
MPLS
IP Service
Tunnel
Security
Quality of Service (QoS)
High Availability
Management & Monitoring
Marvell® Teralynx® 7
(IVM777xx / IVM775xx / IVM773xx)
The Teralynx 7 family is Marvell’s programmable Ethernet switch silicon scaling from 3.2 Tbps to 12.8 Tbps, covering ToR, leaf, and high-radix spine deployments. Built for cloud, enterprise, HPC, and service provider data centers, it combines large buffers, ultra-low latency, and line-rate programmability with industry-leading power efficiency — 2× performance-per-watt over alternatives. Open SDK and SAI APIs enable native SONiC support with a common software stack shared across the entire Teralynx product line.
Architecture & performance
- TeraSCALE™ switch and buffer fabric — scalable from 3.2 Tbps to 12.8 Tbps within a unified architecture, enabling ToR, leaf, and spine in a single silicon family
- Large shared packet buffer with dynamic thresholds for superior microburst (µBurst) absorption under heavy east-west load Cut-through latency of ~500 ns; industry-leading 2× performance-per-watt advantage over alternatives Up to 128 × 100G, 64 × 200G, or 32 × 400G single-chip configurations; multi-chip chassis scales to 256 × 400G (1024 × 100G)
SerDes & port flexibility
- IVM77700 / IVM77610: 50G PAM4 SerDes — industry’s first high-port-count 200G and 400G connectivity; supports 10/25/40/50/100/200/400 GbE
- IVM77500 / IVM77310: 25G NRZ SerDes — supports 10/25/40/50/100 GbE with full backward compatibility
- Up to 256 SerDes (IVM77700/IVM77500); proven robust interoperability across all speed grades
Programmable forwarding
- InnoFlex™ line-rate programmable forwarding pipeline — extend new protocols via software upgrade, no ASIC respin required
- Comprehensive tunneling: IP-in-IP, GRE, MPLS, VXLAN, and Geneve; highly scalable IPv4/IPv6 L2/L3 forwarding tables
- Advanced QoS and traffic management: DCB, RoCEv2, and QCN for lossless RDMA, storage, and AI/ML traffic
Telemetry & openness
- FLASHLIGHT™ visibility and analytics engine with P4 in-band network telemetry (INT) and critical extensions
- Predictive analytics and automated remediation for self-healing networks, reduced OPEX, and improved uptime
- Open SDK + OCP SAI APIs with native SONiC support; common software stack shared with Teralynx 10 for seamless migration
| Low-latency AI RoCE Switch family datasheet | Low-latency AI RoCE Switches Datasheet (PDF) |
| Quick start guide for AsterNOS CLI | AsterNOS Documentation |
The dual 1+1 hot-swappable power supplies ensure continuous operation by allowing replacement without system shutdown.
Equipped with 4+1 hot-swappable fan modules, the system delivers reliable cooling performance, maintaining stability and efficiency under demanding workloads.
| CERTIFICATE | ||||||
| SCOPE | Quality management system | Environmental management | Occupational health and safety | Information security management | US radio frequency compliance | EU health, safety and environmental compliance |
