Open Networking Archives

State of Open Networking 2024: From Niche to Mainstream

Open networking entered 2024 as a mainstream infrastructure choice. The days of open networking being a niche option for technology-forward operators willing to tolerate rough edges are over. The market has crossed a structural threshold: the combination of hardware maturity, NOS feature depth, ecosystem breadth, and production deployment scale has made disaggregated networking the default rational choice for a growing majority of new network infrastructure decisions.

By the Numbers: OcNOS in 2024

Metric	Value
Customers worldwide	600+
Active deployments	10,000+
Countries	40+
Validated hardware platforms	100+
Years of production deployment	25+ (OcNOS CP / ZebOS lineage)
OcNOS 6.5 new features	OpenConfig YANG, 400G spine platforms, Flex-Algo foundations

AI Data Centers: The New Adoption Driver

The most significant new driver for open networking in 2024 is AI data center infrastructure. GPU clusters require switching fabric at scales and port densities that were previously achievable only with the most expensive proprietary platforms. The arrival of Broadcom Tomahawk 4 and Tomahawk 5 on ONIE-capable white-box hardware — running OcNOS with full PFC/ETS lossless fabric support — gives AI infrastructure builders a path to high-performance GPU interconnects at 40–60% lower hardware cost than proprietary alternatives.

The Protocol Stack Is Complete

A common proprietary vendor argument against open networking was protocol immaturity. By 2024, this argument no longer holds for OcNOS:

SR-MPLS with TI-LFA: production-proven across hundreds of SP deployments
EVPN (all service types): E-LINE, E-LAN, E-TREE, L3VPN, IRB — complete
Flex-Algo: production network slicing for 5G and enterprise traffic separation
OpenConfig YANG via NETCONF: vendor-neutral automation fully supported
gNMI on-change telemetry: real-time streaming replacing SNMP polling
IPoDWDM: coherent optical management from OcNOS CLI

What’s Next: OcNOS 6.5 and Beyond

OcNOS 6.5 (August 2024) marks the arrival of OpenConfig YANG model support and Flex-Algo production deployment. OcNOS 6.6 (March 2025) delivers full production Flex-Algo with delay metrics and AI fabric PFC/ETS. OcNOS 7.0 (March 2026) brings 800G AI fabric on Tomahawk 5, container lifecycle management, RPKI peering security, and on-change gNMI telemetry across the full platform portfolio.

IP Infusion Marketing Team

A Practical Guide to Network Disaggregation: Lessons from Real Deployments

Theory about open networking is abundant. Practical guidance from people who have actually done it is rarer. This guide draws on IP Infusion’s experience across 600+ customer deployments to give operators a grounded, actionable framework for evaluating and executing a disaggregated networking deployment.

Question 1: Are Your Use Cases Actually Supported?

The most important question before any disaggregation project is protocol coverage. OcNOS covers the full SP and DC stack, but the specific features you need should be verified against the OcNOS Feature Matrix and tested in the Demo VM before any procurement decision.

Common use cases that are fully supported and production-proven in OcNOS:

BGP full table peering (1M+ routes) with FlowSpec DDoS mitigation
SR-MPLS with TI-LFA fast reroute and Flex-Algo network slicing
EVPN all service types (E-LINE, E-LAN, E-TREE, L3VPN) over MPLS
EVPN-VXLAN data center fabric with PFC/ETS lossless transport
IEEE 1588v2 PTP for 5G timing with hardware timestamping
IPoDWDM with 100G/400G coherent optics managed via OcNOS CLI

Question 2: Hardware Selection

OcNOS runs on 100+ validated platforms. The right hardware depends on:

Factor	Guidance
Port speed and density	Match silicon generation to your capacity requirements; don’t over-spec
Form factor	Fanless options available for outdoor/remote; standard 1U for rack environments
Silicon vendor	Broadcom for DC (TH5 for AI fabric); Marvell or Broadcom for SP
PTP requirements	Verify hardware timestamping support for 5G timing deployments
Transceiver ecosystem	OcNOS supports third-party optics; verify specific transceivers needed

The Deployment Pattern That Works

The most successful OcNOS deployments follow a consistent pattern:

Lab validation first — download the OcNOS Demo VM, build your topology in GNS3 or ContainerLab, validate your specific configurations. This identifies any feature gaps or configuration quirks before hardware arrives.
Pilot on a non-critical segment — deploy 2–4 OcNOS nodes on a lower-risk part of the network first (a new broadband aggregation point, a secondary peering location) to build operational confidence before touching the core.
Parallel deployment for migration — when replacing legacy hardware, bring OcNOS nodes up in parallel and migrate traffic gradually rather than doing a single cutover.
Automation from day one — even for small deployments, invest in Ansible or NETCONF templates early. The operational efficiency at scale depends on consistent, automated configuration management.

What Changes Operationally

The biggest operational adjustment for teams coming from proprietary NOS is not the CLI — OcNOS CLI is industry-standard and familiar. It is the hardware independence. When something goes wrong, the diagnostic divide between hardware and software is cleaner — you can swap hardware without affecting software state, and you can update software without hardware maintenance windows. This is a workflow change that teams adapt to quickly once they experience it.

IP Infusion Marketing Team

Why Service Providers Are Deploying Open Networking: 5 Drivers Behind the Shift

In just over one year, more than 100 of IP Infusion’s service provider customers deployed open networking across their production networks. This is not a pilot or proof-of-concept trend — these are live network deployments carrying commercial traffic. Understanding why SPs are moving now — and moving quickly — reveals the structural forces reshaping the carrier equipment market.

Driver 1: 5G Infrastructure Economics

5G deployment requires a massive expansion of the transport and backhaul network. Every cell site needs a cell site router. Every region needs aggregation capacity. Building this infrastructure with traditional proprietary equipment at traditional prices is economically challenging for most operators outside the Tier-1 carriers.

Disaggregated cell site routers running OcNOS cost 50–65% less than equivalent proprietary platforms. At scale — hundreds or thousands of cell sites — this difference funds substantial 5G rollout acceleration.

Driver 2: Broadband Access Expansion (BEAD and Similar Programs)

The US BEAD program (Broadband Equity, Access, and Deployment) allocated $42.5 billion for broadband infrastructure. Similar programs exist in Europe, Asia, and Latin America. These programs require rapid infrastructure deployment at competitive cost — exactly the value proposition of disaggregated networking. ISPs and electric co-ops building out fiber access networks are adopting OcNOS-based aggregation platforms as a cost-effective foundation for BEAD-funded builds.

Driver 3: Legacy Platform End-of-Life

A wave of proprietary platform end-of-life announcements — Juniper MX204, legacy Cisco ME-series, various white-box platforms with discontinued support — forced operators into a replacement decision. Operators who had previously evaluated open networking but deferred adoption are now actively migrating rather than re-entering another proprietary vendor cycle.

Driver 4: Supply Chain Diversification

The 2020–2023 semiconductor shortage made single-vendor dependency a business risk, not just a procurement preference. Service providers with disaggregated architectures maintained deployment schedules by shifting to alternative ODMs. Those locked into single-vendor procurement faced 12–24 month equipment delays. This experience accelerated multi-vendor sourcing strategies that benefit disaggregated architectures.

Driver 5: Operational Maturity of Open NOS

Five years ago, a common objection to open networking was that the software was not mature enough for production SP use cases. OcNOS 7.0 — with TL 9000 certification, MEF 3.0 compliance, O-RAN validation, 600+ customers, and 10,000+ production deployments — has eliminated that objection. The maturity argument has reversed: operators now ask what risk there is in not adopting open networking while competitors reduce their infrastructure costs.

IP Infusion Marketing Team

End-to-End Disaggregated Networks with OcNOS: From Access to Core

One of the most operationally compelling aspects of OcNOS is its breadth across network segments. Rather than running different NOS platforms at different layers — a common legacy pattern that creates operational silos, separate training requirements, and inconsistent tooling — OcNOS provides a single platform with consistent CLI, management interfaces, and automation tooling from the access edge to the core.

OcNOS Across the Service Provider Network

OcNOS end-to-end: a single network operating system spanning cell site routers, access routers, aggregation, core/transport, IPoDWDM optical, and data center interconnect — with consistent CLI, automation interfaces, and management tooling across all segments.

The Operational Advantage of a Single NOS

Running a single NOS across the full network stack delivers compounding operational benefits:

Consistent CLI — engineers trained on OcNOS can operate any segment of the network without learning a new command syntax
Unified automation — Ansible playbooks, NETCONF templates, and gNMI subscriptions work identically across CSR, aggregation, and DC platforms
Single TAC contact — one support organization owns all issues regardless of which segment they originate in
Consistent upgrade path — OcNOS releases ship simultaneously for SP and DC, with a single upgrade qualification process

IP Infusion Marketing Team

The Business Case for Network Disaggregation: TCO, Flexibility, and Innovation Velocity

The decision to adopt disaggregated networking is not just a technology choice — it is a business decision with measurable financial and operational consequences. This article presents the business case across the dimensions that matter most to operators: total cost of ownership, flexibility, supply chain resilience, and innovation speed.

Total Cost of Ownership: Beyond Hardware Price

The most visible cost difference between proprietary and disaggregated networking is hardware price. White-box platforms from ODMs running on commodity merchant silicon typically cost 50–70% less than equivalent branded hardware. But this is only part of the TCO story.

Cost Factor	Proprietary	Disaggregated (OcNOS)
Hardware (Year 1)	Baseline	50–70% lower
Software licensing	Per-feature or bundle; annual renewal	All-inclusive per-platform; predictable
Support contracts	Tied to hardware vendor; premium pricing	IP Infusion TAC; competitive
Hardware refresh	Forced replacement at EoL; same vendor	Replace hardware only; NOS continues
Transceiver costs	Often proprietary optics required	Third-party optics supported
Training	Vendor-specific; high cost	Industry-standard CLI; minimal delta
Supply chain optionality	Single vendor; exposure to shortage	Multiple ODM sources; flexible

Over a 5-year lifecycle, the cumulative TCO advantage of disaggregation typically reaches 40–60% compared to equivalent proprietary deployments. The hardware savings compound with software flexibility — when a platform reaches end-of-life, operators replace only the hardware and continue running the same OcNOS version, avoiding forced software migrations.

Hardware Refresh Flexibility

One of the most underappreciated benefits of disaggregation is the ability to refresh hardware without changing NOS. In a proprietary network, a platform end-of-life forces both hardware and software replacement simultaneously — a high-risk, high-cost event. In a disaggregated network, hardware refresh is a modular operation:

Replace only the hardware reaching end-of-life
OcNOS configuration migrates directly to the new platform
Engineers work in the same CLI they already know
New silicon generations (400G, 800G) are adopted by upgrading hardware only

Supply Chain Resilience

The 2020–2023 semiconductor shortage exposed a structural risk in single-vendor networking procurement. Operators who standardized entirely on one proprietary vendor faced multi-year wait times for hardware. Disaggregated operators with multi-ODM flexibility could shift procurement to whichever ODM had inventory available — without changing their NOS or retraining their operations teams.

IP Infusion supports 100+ validated hardware platforms across UfiSpace, Edgecore, Celestica, Accton, and others. When one ODM faces supply constraints, operators shift procurement to an alternative — same OcNOS, different hardware.

Innovation Velocity

Proprietary NOS vendors release major versions on 12–24 month cycles, driven by their own roadmap priorities. OcNOS follows a customer-driven roadmap with quarterly releases. Service providers and data center operators who are IP Infusion customers directly influence the feature roadmap through structured feedback channels — a structural advantage that proprietary vendors cannot match.

IP Infusion Marketing Team

What Is Network Disaggregation? A Plain-Language Guide

Network disaggregation is one of the most significant structural shifts in the networking industry in decades. But the term is used so broadly — sometimes to mean open source software, sometimes white-box hardware, sometimes SDN — that its practical meaning gets lost. This article explains exactly what network disaggregation is, what it is not, and why it matters for operators evaluating their infrastructure choices today.

The Traditional Bundled Model

For most of networking’s history, hardware and software were sold as an inseparable bundle. A Cisco router ran Cisco IOS. A Juniper switch ran Junos. The hardware and software were engineered together, tested together, supported together, and priced together. Operators had no choice about which software ran on which hardware — the vendor made that decision.

This model worked well when networking hardware was specialized and proprietary. But it created structural problems as the industry matured:

Vendor lock-in — once you standardized on a vendor’s hardware, you were committed to their software, support pricing, and roadmap priorities
Hardware dependency — when a platform reached end-of-life, operators faced a forced hardware replacement even if the software was working perfectly
Pricing leverage — captive customers have limited negotiating power at renewal time
Innovation pace — vendor roadmaps moved at the vendor’s speed, not the operator’s

What Disaggregation Changes

Network disaggregation separates the networking software from the hardware, allowing each to be sourced, evaluated, and replaced independently. This is exactly what happened in the server market in the 1990s: x86 servers from Dell, HP, or Supermicro all ran the same Linux operating system. The compute market disaggregated, and the result was an explosion of competition, price reduction, and innovation.

Disaggregated Model OcNOS (Network Operating System) Open, carrier-grade NOS — choose independently UfiSpace white-box Edgecore white-box Celestica white-box ✓ Freedom: choose any validated hardware

The fundamental shift from bundled to disaggregated networking. In the disaggregated model, OcNOS runs on any ONIE-capable white-box hardware from multiple ODMs — operators choose hardware and software independently based on price, performance, and availability.

The Three Layers of Disaggregation

Full network disaggregation operates at three levels:

1. Hardware Disaggregation

Open, ONIE-enabled white-box hardware from ODMs (Original Design Manufacturers) like UfiSpace, Edgecore, and Celestica. These platforms run commodity merchant silicon (Broadcom, Marvell) and cost 50–70% less than equivalent branded alternatives. ONIE (Open Network Install Environment) is the standard bootloader that allows any compliant NOS to be installed on any ONIE-capable platform.

2. Software Disaggregation

A carrier-grade network operating system that runs independently of any specific hardware vendor. OcNOS from IP Infusion is the most widely deployed commercial NOS in this category — 600+ customers, 10,000+ active deployments across SP and DC networks worldwide.

3. Management Disaggregation

Centralized management and automation using open standards — NETCONF, gNMI, OpenConfig YANG — rather than vendor-proprietary management systems. IP Maestro provides a GUI-based element management layer for OcNOS networks, while standard interfaces enable integration with existing OSS/BSS systems.

What Disaggregation Is Not

Not DIY — disaggregation does not mean assembling open-source components and engineering your own NOS. OcNOS is a commercial, supported product with a defined roadmap and 24×7 TAC.
Not just cost reduction — while TCO reduction is real and significant, disaggregation also enables hardware refresh flexibility, supply chain resilience, and faster adoption of new silicon generations.
Not limited to data centers — OcNOS runs across access routers, cell site routers, aggregation platforms, core nodes, and optical transport — the full SP network stack.

IP Infusion Marketing Team