Gerben Wierda published another AI-buster article describing what exactly “state-of-the-art” means in AI benchmarks.

Hint: you give an AI model 32 step-by-step examples before asking a question, and it still gets it wrong 10% of the time.

After a brief introduction of how the language models fit into the AI/ML landscape, Javier Antich explained the language model basics, including auto-regression, types of language models, the specifics of large language models, and potential use cases,

TL&DR: If you’re using netlab to build labs for your personal use, you can skip this one, but if you plan to use it to create training labs (like my BGP labs project), you might want to keep reading.

Like any complex enough tool, netlab eventually had to deal with inconsistent version-specific functionality and configuration syntax (OK, topology attributes). I stumbled upon this challenge when I wanted to make labs that use two types of configurable devices.

A previous BGP lab focused on the customer side of BGP communities: adding them to BGP updates to influence upstream ISP behavior. Today’s lab focuses on the ISP side of the equation: using BGP communities in a routing policy to implement RFC 1998-style behavior.

I always said that the Trivia Pursuit certification tests (or job interviews) are nonsense and that one should focus on fundamentals.

In a recent blog post, Daniel Dib described a fantastic scenario: using a simple “why can’t I connect to a web site” question, explore everything from ARP/ND to DNS and TLS.

Obviously, you’ll never see anything that sane in a certification test. An interactive interview doesn’t scale (beyond CCDE), and using humans (and common sense judgment) creates potential legal liabilities (there were rumors that had been one of the reasons a talk with a proctor who could flunk you was dropped from the CCIE test).

Drew Conry-Murray published a excellent summary of his takeaways from the AutoCon0 event, including this one:

Most companies want vendor-supported tools that will actually help them be more efficient, reduce human error, and increase the velocity at which the network team can support new apps and services.

Yeah, that’s nothing new. Most Service Providers wanted vendors to add tons of nerd knobs to their products to adapt them to existing network designs. Obviously, it must be done for free because a vast purchase order¹ is dangling in the air. We’ve seen how well that worked, yet learned nothing from that experience.

Did you find the Network Automation with GitHub Actions blog post interesting? Here are some more GitHub Self-Hosted Runner goodies from Julio Perez: Network CI and Open Source – Welcome to the World of Tomorrow. Enjoy!

Want to explore SRv6? Cisco engineers put together a repository containing scripts and configs for building SRv6 test topologies. It works with Containerlab and FRR (unless you want to beg a Cisco account team for a Cisco 8000 image or make a sandwich while the IOS XRd image is booting).

Want to use netlab? Jeroen van Bemmel implemented baseline SRv6 support for Nokia SR OS.

Did you know that netlab includes full-blown IP address management? You can define address pools (or use predefined ones) and get IPv4 and IPv6 prefixes from those pools assigned to links, interfaces, and loopbacks. You can also assign static prefixes to links, use static IP addresses, interface addresses as an offset within the link subnet, or use unnumbered interfaces.

For an overview of netlab IPAM, watch the netlab address management video (part of the Network Automation Tools webinar), for more details read the netlab addressing tutorial.

On November 22nd, 2023, AMS-IX, one of the largest Internet exchanges in Europe, experienced a significant performance drop lasting more than four hours. While its peak performance is around 10 Tbps, it dropped to about 2.1 Tbps during the outage.

AMS-IX published a very sanitized and diplomatic post-mortem incident summary in which they explained the outage was caused by LACP leakage. That phrase should be a red flag, but let’s dig deeper into the details.