That fancy SDN controller isn't helping if your underlay routing is a mess

I sat in on a troubleshooting call last week where a team spent 3 hours blaming an overlay issue, but the root cause was a simple OSPF neighbor flapping because someone forgot to set a cost on a 10 gig link. Why do people keep tweaking automation scripts before they check the physical routing table?

2 comments

2 Comments

nancy_green26d ago

Fancy tools don't fix broken fundamentals.

rowangonzalez26d ago

You're getting hung up on a single OSPF flap and acting like that proves the whole automation approach is bad. In my experience, the teams that manually check routing tables every time an overlay acts up are the ones burning 3 hours on a call to begin with. With a proper SDN controller and automated underlay validation, you can probe every link in the network in seconds and catch that cost mismatch before it ever becomes a problem. The issue isn't the fancy tool, it's that someone built a brittle automation script that didn't account for basic OSPF config drift. If you tune your automation to monitor OSPF neighbor states and cost values, you'd have an alert before the first user calls in. Your hate for the tools is really a hate for bad implementation, and that's a different conversation.