- The Routing Intent by Leonardo Furtado
- Posts
- If the Control Plane Fails, Nothing Else Matters
If the Control Plane Fails, Nothing Else Matters
Your data plane moves packets. Your control plane moves decisions. If the latter breaks, the former doesn’t matter. This is how we protect the network’s brain.

Data Planes Don’t Think
In hyperscale networks, we take pride in how quickly the data moves. Sub-second failover, lightning-fast ECMP rebalancing, label-switched bypasses; our networks are built to be fast, fluid, and fault-tolerant. And yet, this focus on the data plane often distracts us from a more fragile, more complex, and arguably more critical component: the control plane.
This is where the decisions get made. This is where intent is translated into action. And this is where, if you're not paying attention, everything can fall apart, even when your links are green and your packets are flowing.
You can think of the data plane as muscle: powerful, capable, reactive. But the control plane is the brain. It's the part of the network that holds state, computes reachability, steers flows, and orchestrates policies across thousands, sometimes millions, of endpoints. And like any brain, when it suffers a failure, the rest of the system quickly descends into confusion.
We’ve seen this play out in the real world. Route reflectors silently drop prefixes due to resource exhaustion. A single path computation element goes rogue and floods invalid SRTE instructions. A partial orchestrator outage leads to half-deployed policy states and no clear rollback path. These are battle scars that many large-scale network operators quietly carry.
Yet, even after these hard lessons, many architectures still treat the control plane as an afterthought, something to be deployed with a simple HA pair, or worse, with a single-threaded assumption that if the routers are up, the system must be fine.
But we know better.
This article is about treating the control plane with the respect and rigor it deserves. It’s about engineering high availability, not just for links and line cards, but for protocol state, orchestration pipelines, and the distributed systems that manage reachability at scale. It’s about embracing a new reality:
No matter how fast your packets move, they’re meaningless without the brain that tells them where to go.
Let’s dive in.

Subscribe to our premium content to read the rest.
Become a paying subscriber to get access to this post and other subscriber-only content. No fluff. No marketing slides. Just real engineering, deep insights, and the career momentum you’ve been looking for.
Already a paying subscriber? Sign In.
A subscription gets you:
- • ✅ Exclusive career tools and job prep guidance
- • ✅ Unfiltered breakdowns of protocols, automation, and architecture
- • ✅ Real-world lab scenarios and how to solve them
- • ✅ Hands-on deep dives with annotated configs and diagrams
- • ✅ Priority AMA access — ask me anything