Blog Posts in April 2026
Worth Reading: AI and Knowledge Stagnation
Another week, another interesting AI article (is anyone writing about anything else these days?), this time from Noah Smith (another author worth following). I found this gem hidden in his weekly roundup:
Instead of trying to write a piece of code from scratch, or prove a math theorem from scratch, or figure out some piece of knowledge for yourself, you just ask AI to do it all for you. So everyone ends up getting the right answers to questions whose answers are already known, so they don’t end up adding anything new.
Ten Years of ITNOG
I spent the last two days in Bologna at ITNOG 10 in the excellent company of Italian networking engineers (many of them personal friends) and a few guests from around the world. As always, the organizers and the program committee didn’t disappoint – it was a smoothly organized, lovely event full of interesting presentations. Thanks a million to everyone involved; I’ll definitely be back!
Now for the highlights, starting with the ultimate catnip for the differently attentive: running two presentations in parallel on the same screen with the soundtrack distributed via headphones. I’ve never seen anything like that, and while it looked weird (I have no idea how the presenters took it), it turned out to be very useful, as you could easily tune out AI-washing presentations and switch to something more interesting. On the other hand, you could be faced with a hard choice of having to select one of two excellent presentations:
Hmmm: Rail-Optimized Networking for AI Workloads
Phil Gervasi wrote an interesting article describing Rail-Optimized Networking for AI Training Workloads. Go read it first; I’ll wait.
Does it sound interesting? Were you able to see behind the curtain and figure out what it’s really about?
SR-MPLS Workshop Materials
We’re starting the SR-MPLS workshop @ ITNOG 10 in Bologna in a few hours. Hope you’ll make it, but if you happen to be too far away, here’s the slide deck, the lab topologies, and the usage guidelines.
Lab: Integrated Routing and Bridging (IRB) with EVPN MAC-VRF Instances
Most EVPN documentation rushes into the complexities of symmetric IRB, type-5 EVPN routes, and EVPN IP-VRFs, but if you want to master a technology, it’s better to take it slow and proceed with small steps. In today’s lab exercise, we’ll do just that: explore what happens when you add IP addresses (we’ll use anycast gateways) to VLAN interfaces tied to EVPN MAC-VRF instances.
Testing FRRouting Pull Requests with netlab
Every other blue moon, I discover a bug in FRRouting (example). Because the FRRouting developers care about their work, it usually gets fixed within a few days, often resulting in a “can you test this PR?” question.
It turns out that’s surprisingly easy to do with netlab – here’s the step-by-step procedure (assuming you already have the topology file that reproduced the bug):
BGP Labs: Graceful Degradation for Unsupported Devices
A few weeks ago, I described the changes in the online BGP labs that allow you to use most of the common network operating systems as “external” routers1. However, while we keep improving it, netlab still can’t configure all BGP features on all supported devices (PRs from Nokia and Mikrotik fans would be highly appreciated 😎), which means that it’s possible to configure your environment in a way where some of the more complex labs would simply fail to start.
The limited choice of devices for external routers was always well-documented (example), but if you insisted on using unsupported devices, the lab would fail to start with an error message, and you’d have to tweak the lab topology (example). Wouldn’t it be better to start the lab with a warning?
What is Netlab and Why Network Engineers Should Care
Milan Zapletal, the author of the biggest netlab topology I’ve seen so far (full story), has published a nice introductory article explaining What is Netlab and Why Network Engineers Should Care.
Thank you, Milan! Much appreciated ;)
netlab 26.04: EXOS, BGP Prefix Origination, More Static Routes
netlab release 26.04 is out. Here are the highlights:
- Extreme Networks EXOS is supported as a Vagrant box or containerlab node with OSPF, VLAN, and VRRP configuration (by Seb d’Argoeuves).
- The new bgp.advertise node attribute allows you to advertise networks in the IP routing table into BGP. It’s supported on most platforms.
- The bgp.originate attribute is now dual-stack and VRF-aware, allowing you to originate IPv4 and IPv6 prefixes into per-VRF BGP instances.
- New platforms with static route support: FortiOS (by Aleksey Popov), Nexus OS, Nokia SR OS, Nokia SR Linux. OpenBSD got discard static routes.
Public Videos: Docker 101
While according to the GIFEE True Believers™, Docker is dead and Kubernetes rules the world, people who want to have a bit of life might be perfectly happy running “obsolete” stuff like Docker on their laptops or Linux VMs.
If you happen to be one of the latter, you might like the Introduction to Docker webinar I put together a few years ago. It’s now public; you can watch it with an ipSpace.net account.
Interoperability of EVPN/VXLAN with IPv6 Next Hops
Another chapter in the never-ending "SIP of Networking" saga
FRRouting release 10.6 promised “BGP IPv6 VTEP support,” claiming “it enables EVPN deployments using IPv6 tunnel endpoints while maintaining full backward compatibility with IPv4 VTEPs.” Of course, I had to try it out, and since we already have EVPN over IPv6 running on Arista EOS (since netlab release 26.01), I decided to set up a simple lab with an Arista cEOS device running release 4.35.2F and the latest FRRouting container.
I was not exactly surprised when it did not work. While Arista accepted FRRouting EVPN routes, the FRRouting BGP daemon rejected routes sent by Arista EOS:
Worth Reading: Shameless Guesses, Not Hallucinations
In a recent article, Scott Alexander made an interesting point: What AI produces are not hallucinations but shameless guesses (also known as bullshit) because the training process rewards the correct answers but does not penalize the incorrect ones. After all, having an AI model say, “I don’t know that” is not good for business, is it?
Lab: Summarizing IS-IS Level-1 Routes
IS-IS was designed to carry node addresses (NSAPs) between level-1 routers (called Intermediate Systems) within an area and area prefixes between level-2 routers, resulting in a perfect separation of concerns and forwarding information summarization. When IETF tried to use the same routing protocol for a networking stack with a completely different addressing mentality, something had to give.
… updated on Saturday, April 11, 2026 18:28 +0200
SR Linux Configuration Conversion Tool
A year ago, I was complaining about SR Linux breaking its configuration data model with a new software release. At that time, I was promised it would only happen once a year, and, like clockwork, that moment arrived with the SR Linux release 26.03.
However, this year Miguel Redondo fixed the netlab SR Linux configuration templates (VRF export policies, LocPref routing policy changes) before I could even start looking at them, and Roman Dodin released a tool that tells you exactly what changed between software releases and how to fix it.
Every Layer of Review Makes You 10x Slower
Avery Pennarun published yet another excellent article: every layer of review makes you 10x slower, effectively reiterating what I’ve been saying for decades: all the technology in the world won’t help you unless you re-architect the broken processes.
AI is no exception, but of course, the AI evangelists, LinkedIn AI Wranglers1, and Thought Leaders will never tell you that (or even admit it).