BCM56880 · 1RU 32×400G with IPoDWDM

Broadcom Trident 4 Trident 4 Switches 12.8 Tbps · 32 × 400G · the modern DC leaf.

Two open 1RU 32×400G platforms qualified on OcNOS-DC: Edgecore AS9726-32DB and UfiSpace S9300-32D. The DC-leaf-class chip with IPoDWDM headroom: 400G ZR/ZR+ coherent optics direct in the cage, no transponder or muxponder shelves required.

01The Switches 02Inside the Silicon 03Generation Jump 04OcNOS-DC
Book an Architecture Review Test Drive OcNOS VM
12.8Tbps
Switch Capacity
32×400G
Native Port Radix
1RU
Form Factor
2SKUs
OcNOS-Qualified
400GZR
IPoDWDM Ready
01
The Switches

Two 1RU sibling SKUs. Both on OcNOS-DC.

Both platforms share the same architectural class (32 × QSFP-DD on Broadcom BCM56880), but come from different ODMs. AS9726-32DB ships with ONIE pre-loaded. Both run the same OcNOS-DC software. Choose based on procurement needs and the vendor relationship that suits your fleet. Each card links to the full vendor datasheet (PDF, hosted locally).

Edgecore· DCS series
DC leaf · 400G

AS9726-32DB

Qualified on OcNOS-DC
Ports
32 × QSFP-DD (400G)Breakout: 2×200 / 4×100 / 4×25 / 4×10 (up to 144 logical)
Form
1RU · 438.4 × 590 × 43.5 mm
Power
386 W min. without optics · 1500 W hot-swap redundant PSUs24 W on upper 16 ports · 14 W on lower 16 ports
CPU
Intel x86 CPU · 16 GB DDR4 ECC
▌ Pick this when

400G DC leaf with IPoDWDM headroom: 400G ZR/ZR+ pluggable optics in the upper 16 ports, without transponder or muxponder shelves. Edgecore-branded SKU.

Edgecore AS9726-32DB datasheet PDF
UfiSpace· S9300 series
DC leaf · 400G

S9300-32D

Qualified on OcNOS-DC
Ports
32 × QSFP-DD (400G)Breakout: 4×100 per 400G port
Form
1RU · 436 × 645 × 43.8 mm
Power
403 W typical without optics · 1300 W hot-swap 1+1 PSUs400G ZR/ZR+ support on QSFP-DD ports
CPU
Intel x86 CPU · 32 GB DDR4
▌ Pick this when

Same architectural class as AS9726-32DB. Pick by ODM relationship, BoM economics, or where UfiSpace platforms already dominate the rest of your fleet.

UfiSpace S9300-32D datasheet PDF

· How TD4 fits next to the rest of the OcNOS portfolio

vs Tomahawk 4TD4 is half the capacity in half the rack space: DC leaf class. TH4 (25.6 Tbps) is for spine/aggregation roles that need more switching capacity.
vs Tomahawk 5TD4 is 400G; TH5 (51.2 Tbps) is 800G. Pair them: TD4 leaf, TH5 spine, in larger AI fabrics.
vs Trident 3TD3 maxes at 100G ports; TD4 is the modern 400G upgrade path. Drop-in OcNOS-DC compatibility.
IPoDWDM advantage400G ZR/ZR+ coherent optics in the QSFP-DD cages (upper 16 ports on AS9726-32DB) for metro DCI without transponder or muxponder shelves.
02
Inside the Silicon

Trident 4: purpose-built for the DC leaf role.

The BCM56880 Trident 4 is Broadcom's 12.8 Tbps DC-leaf merchant ASIC: half the capacity of Tomahawk 4 at substantially lower per-port cost. Native radix is 32 × 400G (or 64 × 200G, 128 × 100G via breakout). Uses 50G PAM4 SerDes: 256 lanes, 8 lanes per QSFP-DD cage.

What makes TD4 a leaf-specific chip rather than a smaller TH-series: tuned for east-west forwarding rather than spine aggregation. The 132 MB shared on-chip buffer on both platforms, tighter table sizing for the VXLAN/EVPN VTEP role, and lower power envelope per port are deliberate design choices. TD4 is not the right spine for a 16k-GPU cluster, but it is the right leaf for one.

Buffer132 MB SeriesStrataXGS RoleDC Leaf SerDes50G PAM4 · 256 lanes OpticsIPoDWDM ready

· What 32 × 400G looks like

BCM56880 die12.8 Tbps
256 lanes × 50G PAM4 = 12.8 Tbps. Eight lanes per cage → 400G. 1RU box · 400G ZR coherent optics supported.

Why TD4 is the right shape for the modern DC leaf.

Different from TH-series in deliberate ways, each choice optimised for the leaf workload rather than spine capacity.

PRINCIPLE 01

Right-sized for east-west.

12.8 Tbps in 1RU is a natural fit for a DC leaf: it handles 16 × 400G uplinks to the spine plus 64 × 100G downlinks to servers (16 ports via 4× breakout) without overprovisioning silicon for unused capacity.

12.8 Tbps · 1RU
PRINCIPLE 02

IPoDWDM cage budget.

AS9726-32DB supports 400G ZR/ZR+ pluggable coherent optics in its upper 16 ports (24 W each); S9300-32D lists 400G ZR/ZR+ support on its QSFP-DD ports. Plug a coherent module directly into the leaf for metro DCI without transponder or muxponder shelves.

400G ZR/ZR+ ready
PRINCIPLE 03

EVPN-VXLAN at line rate.

Hardware-accelerated VXLAN encap/decap with proper VTEP scaling. ESI-LAG multi-homing on the leaf and full BGP EVPN control plane on OcNOS-DC.

VXLAN VTEP · ESI-LAG
PRINCIPLE 04

Same NOS as the spine.

The same OcNOS-DC software that runs on TH4 and TH5 spines runs on TD4 leaves. One configuration model, one automation surface, one telemetry pipeline across the whole fabric. No leaf-specific OS to maintain.

OcNOS-DC · same software
03
Generation Jump

Four times the capacity and port speed for the same DC leaf role.

TD3-X7 (3.2 Tbps · 32×100G · 25G NRZ) was the common 100G leaf. TD4 quadrupled the spec sheet at the same 1RU footprint. The role didn't change: the modern DC leaf is just bigger.

Switching capacity
3.2 Tbps → 12.8 Tbps

4× capacity at the same 1RU. Same rack space, four times the throughput.

Native port speed
100G QSFP28 → 400G QSFP-DD

4× per-port speed. Same 32-port radix, but each port carries four times more.

Packet buffer
32 MB → 132 MB

AS7726-32X (TD3) to AS9726-32DB and S9300-32D (TD4), per the vendor datasheets.

Coherent optics
SR/LR pluggables → 400G ZR/ZR+

QSFP-DD cage budget hosts coherent. IPoDWDM was a TD4-era arrival.

Continuity: the same OcNOS-DC software runs on TD3 and TD4. Brownfield refresh keeps EVPN configs, BGP peerings, gNMI subscriptions, and Ansible playbooks intact: capacity quadruples, operational model stays. Trident 3 page →
04
What OcNOS-DC Ships

A DC leaf with coherent optics and lossless RoCEv2.

The TD4 leaf gets the same OcNOS-DC feature surface as the spine, plus IPoDWDM machinery for DCI without transponder or muxponder shelves when the leaf doubles as a metro extension point.

BGP EVPN on the leaf.

Full RFC 7432 EVPN control plane on the leaf. ESI-LAG multi-homing for active/active server attachment, mass-withdraw convergence and route-target auto-derivation. The TH spine and the TD leaf run the same OcNOS-DC EVPN software, so they can peer directly.

400G ZR / ZR+ direct in the cage.

Pluggable coherent optics with OIF/CMIS management. No transponder or muxponder shelves required for metro DCI.

PFC + ECN.

Lossless RoCEv2 on the leaf with priority flow control and ECN.

gNMI / OpenConfig.

Per-port counters and PFC pause counts. Plug into Prometheus/Grafana.

BGP · OSPF · IS-IS.

Full routing stack on the leaf. Treat it as a real router, not a flat L2 switch.

Same OcNOS-DC software as the rest of the fabric.

Layer 3 routing · L1/L2 · AI fabric primitives · Multicast · QoS · Security · Hardware · Management. Per-platform qualification visible on the public OcNOS Feature Matrix.

EVPN-VXLAN ESI-LAG RoCEv2 / PFC 400G ZR / ZR+ BGP / OSPF / IS-IS gNMI / NETCONF ZTP
Day-0 to Day-2

ZTP. gNMI on-change. NETCONF + YANG. DCBX.

Bring up a TD4 leaf in the rack with zero-touch provisioning. Stream every counter to your observability stack without glue scripts.

ZTP IPv4/IPv6 gNMI NETCONF OpenConfig YANG CMIS optics Ansible

Three operator profiles on one leaf-class chip.

Same TD4 silicon, three different DC roles, each leveraging different sides of the same chip.

The 400G leaf upgrade from a 100G fabric.

"Our DC leaf is 100G QSFP28 today. Rack densities are climbing. We need 400G to the leaf, but we're not ready to redesign the spine."

TD4 leaves at 32×400G with 4× breakout to 100G servers. Same EVPN-VXLAN config language as the TD3 fabric they replace. Same OcNOS-DC software. The spine doesn't change.

DC · Leaf Refresh

DC Operator · 400G Leaf

Metro extension with no transponder or muxponder shelves.

"We have two DCs 80 km apart. Optical group wants to retire the transponder and muxponder shelves. Network group wants 400G between them. Procurement wants one box, not two layers."

TD4 leaf with 400G ZR pluggables direct in QSFP-DD cages. EVPN inter-DC stretches the L2/L3 fabric. Coherent IPoDWDM replaces the transponder and muxponder shelves. One box, both layers.

DCI · Metro

Metro DCI · IPoDWDM

Single-pod GPU cluster on a single-tier fabric.

"Our cluster is 32 GPU servers at 100G NICs. We don't need a multi-tier Clos. But we want lossless RoCEv2 and the option to grow."

A pair of TD4 switches forms a single-tier fabric for a single-pod cluster with 100G NICs (4× breakout). RoCEv2 lossless runs from server to switch, with ESI-LAG for multi-homed compute. Add a TH4 or TH5 spine to scale to multiple pods.

DC · Small AI Fabric

AI Cluster · Small / Medium

Full Feature Matrix AI Fabric Solution Coherent DCI Deep-Dive Hardware Compatibility List

Questions architects ask

Two open-hardware 1RU 32×400G platforms: Edgecore AS9726-32DB and UfiSpace S9300-32D. Both platforms share the same architectural class (32×QSFP-DD on Broadcom BCM56880), but come from different ODMs. AS9726-32DB ships with ONIE pre-loaded. Both run the same OcNOS-DC software as the TH5 spines and TH4 aggregation switches.
TD4 (BCM56880) is 12.8 Tbps in 1RU with 32×400G, half the capacity of TH4 (BCM56990) which is 25.6 Tbps in 2RU with 64×400G. TD4 is the DC-leaf-class chip: smaller box, lower power, lower cost per port. TH4 is the spine/aggregation-class chip with more switching capacity. Both are shallow shared-buffer StrataXGS switches; use a StrataDNX router such as Jericho or Qumran for gigabyte-scale deep buffering. In a leaf-spine fabric, use TD4 in the leaf and TH4 or TH5 in the spine. The same OcNOS-DC software runs on both.
Yes, if the cluster is small enough. A 12.8 Tbps spine has 32 × 400G ports, providing one 400G uplink each for up to 32 leaves. Above that, use TH4 (25.6 Tbps) or TH5 (51.2 Tbps) for proper Clos scaling. The same OcNOS-DC software runs on all three.
IP-over-DWDM. Both platforms take 400G ZR/ZR+ pluggable coherent optics: AS9726-32DB in its upper 16 ports, S9300-32D on its QSFP-DD ports. Plug a coherent module directly into the leaf: no transponder, muxponder or regenerator shelves (regenerators within 400G ZR/ZR+ reach). For metro DCI between two leaf switches, that is coherent IPoDWDM in place of traditional DWDM. Saves CapEx, OpEx, and rack U.
Yes. TD4 has the same Broadcom shared-buffer architecture and PFC + ECN primitives as TH-series. OcNOS-DC ships PFC, ETS, ECN and PFC Deadlock Detection & Recovery on TD4 platforms. Dynamic load balancing (DLB) is ASIC-supported and configurable. The 12.8 Tbps capacity makes TD4 a good fit for the leaf in a small-to-mid AI fabric: RoCEv2 lossless from server to leaf, then lossless across the spine on TH4 or TH5.
For 800G ports today, use TH5. For a pure 100G/25G DC leaf where the cluster is below ~64 servers, TD3-X7 is much cheaper. For SP edge or carrier core, use Qumran or Jericho for their different feature set. For DCI roles where the box must absorb deep bursts, use a deep-buffer StrataDNX router such as Jericho or Qumran. TD4 fits modern 400G DC leaf roles with IPoDWDM headroom.
No. Broadcom's public part number for the 12.8 Tbps 32×400G Trident 4 is BCM56880 (Trident4-X11); BCM56990 is Tomahawk 4, a different chip family. The silicon shipping in AS9726-32DB and S9300-32D is the BCM56880 Trident 4. Cross-check with the linked Edgecore and UfiSpace datasheets if exact part numbers matter for your procurement.

Go deeper. Take it with you.

The product datasheet plus short, technical downloads that go further than this page.

Datasheet

OcNOS-DC Datasheet

Full OcNOS-DC specification: the EVPN-VXLAN and Ethernet for AI feature set, software SKUs, supported hardware platforms, and the solution ordering guide.

Get the datasheet
Solution Brief

EVPN-VXLAN Data Center Fabric

Leaf-spine data center fabric on EVPN-VXLAN: a VRF and VNIs per tenant, EVPN multihoming, BGP unnumbered and route-target filtering.

Get the brief
Solution Brief

OcNOS 800G Lossless AI Fabric

Lossless RoCEv2 fabric on Tomahawk 5 spines: SKU tiers, qualified platforms, and deployment architecture.

Get the brief

Design the whole AI fabric with OcNOS

From the business case to the port-count maths, pick up wherever you are in the build.

Solution Open AI Fabric The complete 800G AI fabric: open switches, OcNOS-DC, and support under one contract. Reference designs AI Fabric Topologies Rail-optimized two-tier and three-tier fabrics, and coherent DCI, sized in real port counts. Size & build AI Fabric Design Suite Size a GPU fabric: leaf, spine, and super-spine counts with a component and power summary.
New to AI fabrics? Start here What is an AI fabric? What is a GPU fabric? What is lossless Ethernet? What is RDMA? AI networking switches
The technology inside RoCEv2 lossless Rail-optimized network DLB GLB (7.1) Ultra Ethernet Tomahawk 5 Tomahawk 6 at 1.6T Broadcom silicon DCQCN PFC deadlock InfiniBand vs Ethernet RoCE vs InfiniBand AI fabric architecture Coherent DCI
Evaluate the fabric How to choose an AI networking fabric AI fabric proof points Neocloud networking

Designing a 400G DC leaf with IPoDWDM headroom?

30-minute architecture session with an OcNOS network architect. Bring your DC layout, server count, and DCI requirements, and leave with a sized BoM around AS9726-32DB / S9300-32D and a placement plan vs the TH-class spine.

Book an Architecture Review Test Drive OcNOS-DC