OcNOS-DC: the data center NOS.

OcNOS-DC is IP Infusion's data center variant of OcNOS, the network operating system in production with 600+ operators across 60+ countries. It is built for leaf-and-spine fabrics that carry EVPN-VXLAN overlays, lossless RoCEv2 traffic for AI and ML clusters, and 400G or 800G DCI links. It runs on qualified Broadcom-based open switches from Edgecore, UfiSpace and Celestica. IP Infusion supports the software, and the hardware too when you buy the switch as an OcNOS System.

OcNOS-DC supports EVPN-VXLAN with Type-2 MAC/IP routes and ARP/ND learning to limit flooding in the overlay, and EVPN multihoming for dual-homed servers and storage. EVPN-VXLAN runs on the Trident 3 and Trident 4 switches. Routed tenant features (IRB, the anycast gateway and Type-5 IP-prefix routes) vary by switch model; the Feature Matrix lists each one. The control plane is MP-BGP EVPN per RFC 7432 and RFC 8365, so it interoperates with standards-based EVPN implementations.

OcNOS-DC carries Priority Flow Control (PFC) with QoS over Layer 3 and Explicit Congestion Notification (ECN) for RoCEv2, so RDMA traffic classes can be paused on a per-priority basis to avoid GPU-to-GPU packet drops. The DC-PLUS edition adds Dynamic Load Balancing and PFC deadlock detection and recovery. The intent is a lossless fabric for xCCL collectives such as NCCL and RCCL. Real production deployments to date have run on the order of 20 to 200 GPUs, and the design scales further as engagements grow.

OcNOS-DC is sold in three editions. DC-MGMT carries Layer 2 switching and Layer 3 routing (OSPF, IS-IS and BGP) for out-of-band management switches. DC-IPBASE adds EVPN-VXLAN, and DC-PLUS adds Dynamic Load Balancing and streaming telemetry. Features vary by switch model; the Feature Matrix lists each platform and edition.

OcNOS-DC supports 18 open hardware platforms from Edgecore, UfiSpace and Celestica, from 1G management switches and 48x25G ToR leaves up to 64x800G spines. Supported Broadcom silicon includes Tomahawk 5 (51.2T, 800G), Tomahawk 4 (25.6T, 400G), Tomahawk 3, Tomahawk 2, Trident 4, and Trident 3. Every switch runs OcNOS-DC in the edition that fits its role, so operators keep one CLI and one automation toolchain across leaf, spine, and border roles.

OcNOS-DC exposes NETCONF and gNMI northbound, with OpenConfig and YANG data models, so it fits existing Ansible or in-house controller pipelines. Zero Touch Provisioning brings a new leaf or spine into the fabric with its software and day-zero configuration, without console access. The DC-PLUS edition includes streaming telemetry over gNMI for closed-loop assurance.

These are three ways OcNOS-DC scales the same fabric. Scale-up grows a single pod's capacity by raising per-switch bandwidth and port density, for example moving from a 100G fabric to 400G and 800G, so the same leaf-spine-super-spine topology feeds larger, higher-bandwidth compute nodes without adding more pods. Scale-out grows capacity by adding more of the same building blocks, more spine and super-spine switches and leaf pairs, widening the Clos topology horizontally so more racks and leaves connect while every endpoint keeps any-to-any reachability. Scale-across extends beyond a single data center: a coherent interconnect links independent pods at different sites, with BGP and EVPN carrying routes between them. OcNOS-DC covers all three, with the edition that fits each switch's role.