Reading academic articles about Internet-wide routing challenges you might stumble upon valley-free routing – a pretty important concept with applications in WAN and data center routing design.

If you’re interested in the academic discussions, you’ll find a pretty exhaustive list of papers on this topic in the Informative References section of RFC 7908; here’s the over-simplified version.

Following “if you can’t explain it, you don’t understand it” mantra I decided to use blog posts to organize my ideas while preparing my Networking Infrastructure as Code presentation for the Autumn 2018 Building Network Automation Solutions online course. Constructive feedback is highly appreciated.

Let’s start with a simple terminology question: what exactly is Infrastructure as Code that everyone is raving about? Here’s what Wikipedia has to say on the topic:

Evil CCIE left a lengthy comment on one of my blog posts including this interesting observation:

It's always interesting to hear all kind of reasons from people to deploy CLOS fabrics in DC in Enterprise segment typically that I deal with while they mostly don't have clue about why they should be doing it in first place. […] Usually a good justification is DC to support high amount of East-West Traffic....but really? […] Ask them if they even have any benchmarks or tools to measure that in first place :)

What he wrote proves that most networking practitioners never move beyond regurgitating vendor marketing (because that’s so much easier than making the first step toward becoming an engineer by figuring out how technology really works).

In recent Software Gone Wild episodes we explored emerging routing protocols trying to address the specific needs of highly-meshed data center fabrics – RIFT and OpenFabric. In Episode 92 with Dinesh Dutt we decided to revisit the basics trying to answer a seemingly simple question: do we really need new routing protocols?

One of my subscribers wondered whether it would make sense to build a traditional leaf-and-spine fabric or go for Cisco ACI. He started his email with:

One option is a "standalone" Spine/Leaf VXLAN-with EVPN deployment based on Nexus equipment. This approach could probably be accompanied by some kind of automation like Ansible to ease operation/maintenance of the network.

This is what I would do these days if the customer feels comfortable investing at least the minimum amount of work into an automation solution. Having simpler technology + well-understood automation solution is (in my biased opinion) better than having a complex black box.

As promised, here’s the second part of my Benefits of Network Automation interview with Christoph Jaggi published in German on Inside-IT last Friday (part 1 is here).

What are some of the challenges?

The biggest challenge everyone faces when starting the network automation is the snowflake nature of most enterprise networks and the million one-off exceptions we had to make in the past to cope with badly-designed applications or unrealistic user requirements. Remember: you cannot automate what you cannot describe in enough details.

I had a great chat about the benefits of network automation with Christoph Jaggi a while ago, resulting in 2-part interview published by Inside-IT. As you might prefer to read the English original instead of using Google Translate, here it is (or you could practice your language skills and read the German version).

We migrated the self-study materials for the network infrastructure and services module of the Building Next-Generation Data Centers online course into the new format, and split the largest module of the course into manageable chunks: data center fabrics 101, designing leaf-and-spine fabrics, overlay virtual networking, IPv6 and network services.

Feedback on the new format is obviously highly welcome. Thank you!

Christoph Jaggi, the author of Transport and Network Security Primer and Ethernet Encryption webinars published a high-level introductory article in Inside-IT online magazine describing security deficiencies of SD-WAN solutions based on the work he did analyzing them for a large multinational corporation.

As the topic might be interesting to a wider audience, I asked him to translate the article into English. Here it is…

I stumbled upon an article with an interesting title (and worth reading): To Make Self-Driving Cars Safe, We Also Need Better Roads and Infrastructure… and thought about the claims along the lines of “if they managed to solve the self-driving cars challenge, it’s realistic to expect self-driving networks” made in Self-Driving Networks podcast episode. Turns out the self-driving cars problem is far far away from being solved.

The summer break is over, and we’ve already scheduled a half-dozen events and webinars in August and September:

• We’ll start with Introduction to Software-Defined Data Centers webinar on August 28^th.
• There are still a few places left in the Lean Start in Network Automation workshop in Zurich, Switzerland, on August 30^th. Register here.

We’ll run an event or webinar in almost every single week in September:

Tom Hollingsworth published a great blog post summarizing Cloud Field Day presentation by Ben Sigelman.

TL&DR: You’re not Google, you don’t have their problems, and so you’re probably not a good match for their tools.

While this shouldn’t come as a surprise to regular readers of my blog (here’s what I wrote on the topic in 2016), it’s refreshing to see it spelled out so eloquently (and by an ex-Googler).

This blog post was initially sent to the subscribers of my SDN and Network Automation mailing list. Subscribe here.

Tom Limoncelli published a must-read article in ACM Queue describing GitOps – the idea of using Pull Requests together with CI/CD pipeline to give your users the ability to request changes to infrastructure configuration.

Using GitOps in networking is nothing new – Leslie Carr talked about this concept almost three years ago @ RIPE 71, and I described some of the workflows you could use in Network Automation 101 webinar.

When I started the Building Next-Generation Data Centers online course, I didn’t have the automated infrastructure to support it, so I had to go with the next best solution: a reasonably-flexible Content Management System, and Mediawiki turned out to be a pretty good option.

In the meantime, we developed a full-blown course support system, included guided self-paced study (available with most ipSpace.net online course), and progress tracking. It was time to migrate the data center material into the same format.

A while ago I stumbled upon Schneier’s law (must-read):

Any person can invent a security system so clever that she or he can't think of how to break it.

I’m pretty sure there’s a networking equivalent:

Any person can create a clever network design that is so complex that she or he can't figure out how it will fail in production.

I know I’ve been there with my early OSPF network designs.