Why Network Engineers Still Meet in Person
The industry that made remote connection possible still can't shake the conference room, and the reasons why don't all hold up under close examination.
The internet doesn't route around damage on its own. Someone is awake, staring at a graph, deciding what to touch.
The industry that made remote connection possible still can't shake the conference room, and the reasons why don't all hold up under close examination.
Border Gateway Protocol quietly decides where your internet traffic goes, and for decades it's run almost entirely on the assumption that every network on earth is telling the truth. Cryptographic verification is finally catching up, slowly.
Undersea cables get damaged constantly, mostly by anchors and fishing gear, and almost nobody notices because traffic reroutes across other cables and paths. Here's why that's true, and why a few unlucky regions feel every single cut.
A pattern I keep noticing on program committees: the talks that nearly get rejected on a weak abstract are sometimes the best ones in the room. That says more about how we judge submissions than about the submitters.
Moderation queues, bounce storms, and a public archive that still beats every flashier alternative: an honest look at what it takes to keep a network operator mailing list alive in 2026, and why nothing has managed to replace it.
A plain account of why I started Keep Ukraine Connected, what running it for years actually felt like, and why handing it off this year was harder than I expected.
NOG communities run on volunteers stacking event work on top of day jobs, and the warning signs of burnout are almost always quiet until someone is gone. A reflection on why handoff so rarely works, and why the community rarely says so out loud.
Vendor neutrality isn't a mission-page value. It's a decision about who controls the agenda and the badge, and that decision is what makes honest failure talks possible.
Practical, first-person notes on what separates a call-for-papers abstract that gets accepted from one that quietly gets skipped, drawn from years of reading NOG submissions and building the review tooling behind that process.