Boring network design is successful network design: lessons from NANOG 96
In networking there is constant tension: the drive to adopt the newest, most sophisticated technology versus the discipline to keep things simple and predictable. At NANOG 96, Ryan Hamel, network automation engineer at Zayo Group, and Ethan Banks of Packet Pushers, laid out a thesis many experienced operators will recognize immediately: the most successful network is the most boring one.
The ambitious network engineer’s paradox
Talented network engineers naturally solve problems with technically elegant solutions. That is good. When that instinct is over-applied to operational infrastructure, it produces networks that look brilliant on paper and are hell in production.
A network that is hard to understand is hard to operate. A network that is hard to operate fails more often. A network that fails more often has real cost: in engineering time, incidents, and affected customers.
Simplicity is not lack of ambition. It is operational discipline.
What “boring design” really means
Hamel and Banks are not arguing for primitive networks. They argue for networks where:
- Design decisions are predictable: someone joining the team can understand the architecture without weeks of technical archaeology.
- Standardization beats local optimization: same gear, same baseline configuration, same policies at every site.
- Changes have minimal blast radius: if you change something at one site, impact is predictable because all sites look alike.
- Automation is possible: you cannot automate what you cannot predict.
In practical terms: if every router in your network has a different configuration because “that site had special requirements,” you have accumulated technical debt that will be paid at the worst possible moment.
The hidden win: time to think
A benefit rarely mentioned in standardization debates: when the network behaves predictably and operators are not constantly firefighting, they gain time to think.
Hamel and Banks cited Harvard Business Review and Mayo Clinic on the cognitive benefits of productive boredom. It may sound odd in a technical talk, but the point holds: teams running complex, unpredictable networks stay perpetually reactive. There is no room for continuous improvement, training, or strategic innovation.
A boring network gives that space back.
Why this matters especially in Latin America
For ISPs and network companies in Latin America, the pressure to keep things simple is, if anything, stronger:
-
Smaller teams: most regional ISPs do not have Tier-1 staffing. When something fails at 3 a.m., maybe only one person is available. That person needs a network they can understand.
-
Limited training budgets: standardization lowers training cost. You do not need specialists in ten different technologies; you need strong engineers who master one well-chosen stack.
-
High technical turnover: when the engineer who designed the network leaves, how much tacit knowledge walks out with them? Documentation helps, but not always. Standardization makes the architecture self-explanatory.
The boring-network checklist
To apply these principles today:
- Audit variability: how many different baseline configuration versions exist on your devices? Each variant is technical debt.
- Define a golden template per device type: PE router, CE router, access switch, distribution switch—each with a canonical configuration.
- Document exceptions: if a site must differ from standard, document why. When someone needs to understand it later, the “why” matters more than the “what.”
- Resist the elegant solution: if you can solve a problem with BGP communities and scripting, ask whether the same outcome could be achieved with simpler configuration, even if less “elegant.”
- Automate from day one: automation forces standardization. If you cannot automate something, it is not standardized enough.
Consulting can be boring too
At Ayuda.LA we apply this principle in every project. When we design a network for an ISP or enterprise, we do not chase the most sophisticated architecture; we aim for what fits the team that will run it and the years ahead. That sometimes means recommending something less glamorous than the latest market buzz.
The most successful network we have ever seen never appears in a vendor case study. It is the one that has run for years without anyone noticing—because it just works.
Want to assess whether your network infrastructure is ready for the next level of standardization and automation? At Ayuda.LA we run network architecture assessments for ISPs and companies across Latin America. Let’s talk.