Another full day attending a company offsite meeting, but managed to read some insightful articles on platform engineering, machine learning, and feature flags.
[blog] Don’t Call It A Platform. Good rant! Sam looks at the hype around platform engineering and makes an argument that it’s really about developers. That doesn’t seem entirely right—as platforms are more than just destinations for custom-built apps—but I get the focus on catering to the users.
[article] IRS not capturing the true cost of its legacy IT footprint, watchdog says. This US government agency is behind on its modernization plans because modernization is hard. And, apparently no high-value assets are in the cloud yet.
[blog] Google Research, 2022 & beyond: Algorithms for efficient deep learning. Another link-filled post from our research team that explains some of the fundamentals powering today’s best machine learning.
[article] The Dos and Don‘ts of API Monetization. Some of you build APIs that you want to make money from. This post offers some advice for you.
[blog] Application security with Cloud SQL IAM database authentication. Categorize this under “make it easy to do the right thing.” Get yourself out of storing database passwords, and use integrated authentication instead.
[article] Google Touts Web-Based Machine Learning with TensorFlow.js. Don’t have a ton of Python skills or GPUs sitting around? Maybe you should look at browser-based ML model development with something like Tensorflow.js. I’m going to give it a look myself.
[blog] POSIX Compatibility Comparison among four file system on the cloud. If you’re moving more traditional workloads to the cloud and they depend on file system standards like POSIX, make sure you know where to find the most compatibility.
[blog] Feature Flags and How to Use Them in Software Development. Even if you could technically ship software at any moment, your other business teams might not want to expose the new functionality right away. This is a good post about feature flags.
[article] Containers and Serverless—Rivals or Cohorts? To me, the positioning is wrong in this piece. When some people say “containers” they mean “orchestrator” like Kubernetes. But containers are also part of popular serverless platforms. Let’s separate the argument about packaging (e.g. container) from the runtime!
##
Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:
Richard Seroter's Architecture Musings 
One thought