A unified data foundation for a global, multi-metro network

The challenge

Years of growth and acquisition are kind to a balance sheet and hard on an operations team. This provider had inherited a sprawl of network environments, each with its own conventions and its own idea of the truth. IP address management, to take the most painful example, ran on a homegrown database that had quietly drifted out of step with the network below it. DNS, authentication, and time were configured one way in one metro and another way in the next. There was no single source of truth anyone fully trusted, and the human effort to keep it all stitched together grew with every new site. That kind of drift is invisible right up until the day it causes an outage, and then it’s all anyone can talk about.

For a provider whose product is connectivity itself, this is not a back-office nuisance but operational risk under everything the business sells. Worse, it quietly blocks progress on every other front: you cannot automate against data you do not trust, and you cannot modernize on a foundation that argues with itself.

Our role

The client ran a broad, multi-year program, organized under four pillars, to standardize, integrate, govern, and modernize its global network operations. We led the systems-integration and observability workstream. We supported the governance and operations effort, and we advised on the emerging AI layer. The network redesign itself was delivered by the client’s internal teams. Our focus was the data and services foundation the rest of the operation relies on.

What we did

We started where the drift started: with the data. From there we worked outward to the services that depend on it.

• Stood up NetBox as the authoritative source of truth. That was as much archaeology as engineering. We migrated IP address management off the homegrown database and modeled the data the way it should have been. Driving the import meant cleaning up years of accumulated drift so the source of truth actually matched the network.
• Put data stewardship in place so the source of truth stays accurate over time. Clear ownership, conventions, and process, not just a one-time load.
• Built a NetBox-driven DNS pipeline so records are generated from the source of truth instead of maintained by hand.
• Unified authentication (AAA) across the estate on HPE Aruba ClearPass.
• Helped build a global NTP time network, synchronized consistently across metros.
• Tied it together with a solid data and ETL layer that keeps the source of truth and the services it feeds in step.

The order mattered as much as the parts. We fixed the data first, then built the services that draw from it. Each layer we added strengthened the foundation instead of papering over the cracks. We led this work and delivered it hands-on alongside the client’s teams. This was one workstream inside a much larger program.

Why it mattered

A trustworthy source of truth is the thing everything else quietly depends on. Once the data was right and the services drew from it automatically, the operator could change things faster and with less fear. It could automate against numbers it actually trusted and give the wider program’s governance and AI ambitions a foundation the rest of the work could rely on. The unglamorous work at the bottom is what makes the ambitious work at the top possible.

The range it took

Work like this doesn’t sit at a single altitude, and that’s where we do our best work. Over the program we moved between altitudes freely. We set the target architecture, then orchestrated the vendor partners whose products had to fit together. We deployed and integrated them ourselves at the engineering level. Then we rose back up to turn what we were learning on the ground into business strategy leadership could act on.

That range is the whole point. An architecture only holds if someone can carry it all the way down to a working DNS pipeline and a clean AAA cutover, then back up to a straight conversation about risk, cost, and where to invest next. Being fluent with the vendors and credible with the engineers, while staying legible to the executives, is how a program this broad stays coherent instead of splintering into disconnected projects.