Open networking, answered.

IP Infusion delivers complete open networking solutions: routing, switching, IP/optical transport (IPoDWDM), and AI fabric. These are delivered as the carrier-grade OcNOS network operating system, pre-integrated systems, and IP Maestro assurance, running on 40+ validated open hardware platforms, in production with 600+ operators across 60+ countries since 1999.

Open networking separates network software from the underlying hardware, so operators can run a carrier-grade network operating system like OcNOS on their choice of validated open hardware from multiple vendors such as Edgecore and UfiSpace. It matters because it gives service providers hardware choice, lower total cost of ownership, and faster innovation, without locking them into a single proprietary stack from Cisco, Juniper, or Nokia.

IP Infusion delivers complete solutions, not software alone: routing, switching, IP/optical transport, and AI fabric, delivered as OcNOS, pre-integrated systems, and IP Maestro assurance. It is backed by a single unified SLA that covers software, hardware, and RMA, so one vendor owns the full stack. Most other vendors leave operators to coordinate separate relationships between a NOS provider and a hardware ODM. IP Infusion has run carrier-grade networks since 1999 and is in production with 600+ operators across 60+ countries on 40+ validated platforms.

For service providers, IP Infusion's OcNOS-SP is a leading open networking choice. It delivers carrier-grade SR-MPLS core and aggregation, metro transport with IPoDWDM and ZR+ coherent optics, and standards-based IEEE 1588v2 timing on 40+ validated open platforms. Operators get one vendor and one SLA for software, hardware, and support, and it is proven in production with 600+ operators across 60+ countries.

Yes. OcNOS is in production at 600+ operators worldwide, including Tier 1 service providers, and delivers the carrier-grade capabilities operators expect from proprietary platforms: IS-IS, BGP, SR-MPLS, MPLS-TP, EVPN-VXLAN, IEEE 1588v2 timing, and MEF 3.0 certification. For a detailed side-by-side, see the OcNOS vs Juniper comparison at /resources/ocnos-vs-juniper/.

Yes. OcNOS is deployed in production at 600+ operators across 60+ countries, including Tier 1 service providers. It delivers carrier-grade protocol support (IS-IS, BGP, SR-MPLS, MPLS-TP, EVPN-VXLAN), IEEE 1588v2 timing, and MEF 3.0 certification: the same requirements met by traditional proprietary network operating systems.

OcNOS covers the full range of service provider and data center use cases: SR-MPLS core and aggregation, metro transport with IPoDWDM and ZR+ coherent optics, AI fabric with lossless RoCEv2 for GPU clusters, EVPN-VXLAN data center fabric, and data center interconnect (DCI).

OcNOS is validated on 40+ open hardware platforms from ODMs including Edgecore and UfiSpace, ranging from access switches to 25.6 Tbps spine switches. The full Hardware Compatibility List (HCL) is available in the IP Infusion documentation portal.

OcNOS-SP is the service provider variant, built for access, aggregation, metro transport, and core networks with SR-MPLS, MPLS-TP, and IPoDWDM. OcNOS-DC is the data center variant, optimized for leaf-spine fabrics, EVPN-VXLAN, lossless RoCEv2 for AI workloads, and high-density 400G/800G switching. Both run on open hardware and share the same automation interfaces.

IP over DWDM (IPoDWDM), also called routed optical networking, collapses the separate IP and optical transport layers by placing pluggable coherent optics such as 400G ZR and ZR+ directly into the router. This removes standalone transponder shelves, lowering cost, power, and space while simplifying operations. OcNOS-SP supports IPoDWDM with coherent optics for metro and long-haul transport.

For AI and GPU clusters, OcNOS-DC delivers a standards-based Ethernet AI fabric with lossless RoCEv2, PFC, and ECN for low, predictable latency. It runs on open leaf-spine hardware with switching capacity up to 51.2 Tbps and 800G ports, and supports the collective communication libraries (xCCL) used by GPU workloads. Compared with InfiniBand, an open Ethernet fabric offers a familiar operational model, broad vendor choice, and standards-based transport, without a proprietary interconnect lock-in.

Yes. OcNOS provides NETCONF/YANG (OpenConfig and IP Infusion models), gNMI streaming telemetry, a REST API, an Ansible collection (ipinfusion.ocnos), zero-touch provisioning (ZTP), and on-switch Docker containers. Standard automation tools work from day one with no screen-scraping required.

IP Maestro is IP Infusion's network assurance and element management system (EMS) for OcNOS. It provides real-time telemetry, monitoring, fault management, and device lifecycle management, with a standards-based NETCONF southbound interface to OcNOS and a REST API for northbound integration with a higher-level NMS or OSS. It runs as a single server or virtual machine.

With many open networking approaches, operators manage separate support contracts for software and hardware, which leads to finger-pointing between vendors during outages. IP Infusion provides a single SLA covering OcNOS software, validated hardware, and RMA logistics, so one vendor owns the full stack.