From Proof-of-Concept to Production: scaling and migrating IP infrastructure management

We all know that enterprise processes take time. As time goes by and businesses grow, they become slower and less flexible in adopting new things. This is especially true in the case of IT, where every change affects thousands, if not hundreds of thousands of users.

From Excel to excellent

Even in the age of digital transformation and cloud, legacy IP infrastructure is still reasonably ubiquitous in large-scale corporate environments. And it’s no surprise.

The problem with legacy IP infrastructure isn’t that it doesn’t work at all — it’s that it doesn’t work particularly well. And when there’s still substantial, however diminishing, financial value sitting in those hardware and software components, companies are not in a rush to change.

Additionally, after the Proof-of-Concept is successful, there’s still the project of deploying the solution for real, in production. It can be quite a stark difference seeing chaotic Excel sheets transformed into neatly organized IP address management.

But a Proof-of-Concept is often only a percentage of a percent of what the solution has to manage. And production deployment has no margin for error, or the business suffers.

Up and running in a matter of hours

The problem with most IP infrastructure management solutions is the lack of flexibility. What seems easy enough in a lab environment for a Proof-of-Concept can be a nightmare when it needs to be replicated on a global scale for production.

The chief reason is usually the need for appliances that need to be installed (and later maintained, and also paid for again and again). Appliances don’t scale well. Estimating how many you need, and how to expand in sync with your business needs is a puzzle to be solved perpetually. (And what if you actually need to shrink networks? Appliances don’t scale well, and certainly don’t scale well multiple ways.)

And where the Proof-of-Concept environment is often created empty and new, production has live data that is accessed, used, and modified — every second of every hour of every day.

Growing pains are a fact of life. Or are they?

The good thing about an overlay approach to IP infrastructure management is that they don’t rely on appliances. That alone mitigates most of the deployment headaches.

An overlay solution can be deployed the same way on ten devices as on ten thousand. The variable isn’t appliances to strategize, but automated discovery processes to run. Deployment time can be cut from weeks or months down to hours. Overlays don’t change networks: they pull data from the components unobtrusively. And if there’s an issue, they can be turned off without affecting operations.

The real fact of life is that networks have outgrown static management paradigms. Heterogeneous, dynamically scaling networks need flexible, agile control mechanisms. And even if networks were still mostly static, why would you burn money on unneeded deployment and maintenance costs?