From Operator to Agent Manager: The Real Shift in Network Engineering
I’ve been saying for ten years that network automation was three years away from being the only way to do things. Every year I moved the goalposts, because the data never caught up to the rhetoric. After a decade of scripts, YAML, Ansible, Terraform, and every NetDevOps framework we could throw at the problem, roughly […]
For over a decade, network automation remained a pipe dream for many industry experts. Despite the proliferation of scripts, YAML, Ansible, Terraform, and every NetDevOps framework imaginable, roughly 70% of enterprise networks still rely on manual management. The industry has consistently attributed this lag to a lack of adoption, but author contends that the issue lies elsewhere.
Automating the network required engineers to simultaneously become proficient software developers, a daunting task that the author believes was never realistic. The discipline was built around stability and predictability, asking engineers to ship production-grade software on top of their primary responsibility of maintaining network stability.
The real shift, according to author, is not in the underlying capability of automation, but in the interface. When an engineer provides specifications in plain language and an agent translates them, the engineer's expertise remains at the core. The challenge lies in making the on-ramp to automation easy enough for engineers. The job of network engineering is evolving, not disappearing.
Engineers will manage fleets of agents, each handling documentation, compliance checks, ticket triage, and data gathering. This shift aligns with the way network engineers already operate, troubleshooting in layers and designing for redundancy. Automation should start with read-only tasks, building trust incrementally. Engineers should begin by documenting the network, performing compliance checks, and reconciling data sources without touching production.
As trust grows, agents can be granted more privileges, eventually making routine tickets closed without human intervention. This model of leveraging rather than losing jobs is similar to how accountants transitioned from manual reconciliation to using spreadsheets. The author urges engineers to start with one read-only use case and build agents from there, demonstrating how the role is changing one agent at a time.
Written by urgent.news from DevOps.com's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
