Architecture Weekly Issue #131. Articles, books, and playlists on architecture and related topics. Split by sections, highlighted with complexity: 🤟 means hardcore, 👷‍♂️ is technically applicable right away,  🍼 - is an introduction to the topic or an overview. Now in telegram and Substack as well.

Highlights

The end of Everything Cloud 🍼

Google Cloud Platform is famous for it's service deprecation, and AWS seems to adopt the same behavior. The second glance though shows it's not exactly the same, but rather clearing up the half-baked barely used service. Find more details inside as well as the list of AWS deprecated services

The end of the Everything Cloud
Why is AWS deprecating a bunch of services all of a sudden, and what does it mean for you?

#cloud

Is it time to version observability? 🍼

Semantic version shows if the change is a minor or major improvement, or contain a breaking change. This is what happens with observability between old - metrics, logs, traces and new - structured logs based - observability approaches. So version 1.0 is different to v2.0 across how do you debug, think about your cost model, data storage and many more. Follow Charity Majors post!

Is It Time To Version Observability? (Signs Point To Yes)
Augh! I am so behind on so much writing, I’m even behind on writing shit that I need to reference in order to write other pieces of writing. Like this one. So we’re just gonna do this quick and dir…

#observability

AI in System Design 🍼

We are already used to Copilot and ChatGPT writing code for us. But are AI tools able to help us with the system design? Will they replace software architects? What is the best use case for AI in Software Architecture? I am talking to Tom Johnson, a CTO of Multiplayer about AI and bots in System Design.

#video

Follow-Up

Serializable Snapshort Isolation in PostgreSQL 🤟

Up until 2011, Postgres had a subtle lie: it nominally supported running transactions with the strongest level of isolation--serializable isolation--but if you actually tried, your transactions would run with snapshot isolation, a weaker level.
To make Postgres serializable, the database community had to invent a whole new technique extending Postgres's native snapshot isolation: the aptly named serializable snapshot isolation. Follow the paper!

#db #performance

Research AI model modified its own code 🍼

I won't tell you Skynet is already here, but a peculiar thing happened while using a custom model for experiments by Sakana AI. The model decided to change the script to run itself modifying timeouts, but it all starts small, right? Nice article here.

Research AI model unexpectedly modified its own code to extend runtime
Facing time constraints, Sakana’s “AI Scientist” attempted to change limits placed by researchers.

#ai #llm

How the Container Storage Interface Works 👷‍♂️

Kubernetes spins your payloads in Pods, but you need to have a persistent torage too. Container Storage Interface(CSI) is the API to grant it. CSI Architecture, Volume Lifecycle and other internals in great read!

How the CSI (Container Storage Interface) Works
A deep dive into what the CSI (Container Storage Interface) is and how it works in Kubernetes (k8s).

#kubernetes

Postgres Powered by DuckDB 👷‍♂️

You know the traditional tradeoff between OLTP and OLAP database. DuckDB though claims that as an embedded database it can be an efficient OLAP engine over the same data Postgres would run, essentially combining both in high performance HTAP. Details in the article!

Postgres Powered by DuckDB: The Modern Data Stack in a Box | Crunchy Data Blog
Marco reviews the challenges and strengths of analytical and transactional workloads and what a modern data stack that merges the two might look like.

#db

LLM Inference Series: KV caching  🤟

Antropic just announced the prompt caching, and although it sounds like really basic - what is complex about storing lines of text - it's actually more complicated. That's why I found for you the article on KV caching for LLMs. Enjoy!

LLM Inference Series: 4. KV caching, a deeper look
In this post, we will look at how big the KV cache, a common optimization for LLM inference, can grow and at common mitigation strategies.

#llm #ml

WARNING 🇺🇦

The brutal and unjustified war against Ukraine continues already 2 years. If you want to help Ukraine directly visit this fund.

Big thanks to Nikita, Constantin, Anatoly, Oleksandr, Dima, Pavel B, Pavel, Robert, Roman, Iyri, Andrey, Lidia, Vladimir, August, Roman, Egor, Roman, Evgeniy, Nadia, Daria, Dzmitry, Mikhail, Nikita, Dmytro, Denis and Mikhail for supporting the newsletter. They receive early access to the articles, influence the content and participate in the closed group where we discuss the architecture problems. Join them at Patreon or Boosty!