I had written about this here in the context of Amazon operational excellence

muratbuffalo.blogspot.com/2024/08/loom...

OSTEP Chapter 14: Interlude -- Memory API This is a short chapter covering the nuts and bolts of memory allocation in C: malloc(), free(), and the many ways programmers get them wro...

[new blog post]

OSTEP Chapter 14: Interlude -- Memory API

muratbuffalo.blogspot.com/2026/04/oste...

Immaculate vibes indeed.

PolitePaxos: A New Consensus Variant Where the Proposer Just Asks Nicely Paxos consensus protocol, despite its many theoretical virtues, is fundamentally rude. One need only look at the way it behaves to see the p...

[new blog post]

PolitePaxos: A New Consensus Variant Where the Proposer Just Asks Nicely

muratbuffalo.blogspot.com/2026/04/poli...

the long task gap is real. single-file refactors feel magical, then you ask it to trace a bug across four services and it starts hallucinating import paths that don't exist

Measuring AI Ability to Complete Long Software Tasks This paper from METR (Model Evaluation & Threat Research) introduces a new metric for tracking AI progress: the "50%-task-completion time h...

[new blog post]

Measuring AI Ability to Complete Long Software Tasks

muratbuffalo.blogspot.com/2026/03/meas...

OSTEP Chapter 13: The Abstraction of Address Spaces Chapter 13 of OSTEP provides a primer on how and why modern operating systems abstract physical hardware. This is part of our series going t...

[new blog post]

OSTEP Chapter 13: The Abstraction of Address Spaces

muratbuffalo.blogspot.com/2026/03/oste...

SysMoBench: Evaluating AI on Formally Modeling Complex Real-World Systems This paper presents SysMoBench, a benchmark designed to evaluate generative AI's ability to formally model complex concurrent and distribute...

[new blog post]

SysMoBench: Evaluating AI on Formally Modeling Complex Real-World Systems

muratbuffalo.blogspot.com/2026/03/sysm...

TLA+ mental models In the age of LLMs, syntax is no longer the bottleneck for writing, reading, or learning TLA+. People are even getting value by generating T...

[new blog post]

TLA+ Mental Models

muratbuffalo.blogspot.com/2026/03/tla-...

#tlaplus

Break Paxos As I mentioned in my previous blog post, I recently got my hands on Claude Code. In the morning, I used it to build a  Hybrid Logical Clock...

[new blog post]

Break Paxos

muratbuffalo.blogspot.com/2026/03/brea...

Claude Code experiment: Visualizing Hybrid Logical Clocks Yesterday morning I downloaded Claude Code, and wanted to see what this bad boy can do. What better way to learn how this works than coding ...

[new blog post]

Claude Code experiment: Visualizing Hybrid Logical Clocks

muratbuffalo.blogspot.com/2026/03/clau...

It got the safety logic part first try,

but had a bug in timeout setting, which hilariously caused the leader to timeout and duel with itself, even when it is the leader.

After pointing this out, it fixed it now.

Using Claude Code, I built another distributed systems visualization.

This one is an interactive Paxos tutorial game. You try (and fail) to break safety by killing and delaying nodes. Paxos holds its ground. What’s your high score?

Play it here: muratdem.github.io/BeatPaxos/

Share and enjoy.

You can send messages by drag and drop, create local events by double click, and take a snapshot at T-1, by pressing the snapshot button.

Since then HLC has been adopted by many distributed databases, including MongoDB, CockroachDB, Amazon Aurora DSQL, YugabyteDB, etc.

So, try it here live:
muratdem.github.io/hlc-visualiz...

Hybrid Logical Clocks Here I will write about our recent work on Hybrid Logical Clocks, which provides a feasible alternative to Google's TrueTime. A brief hist...

HLC is a simple idea we proposed in 2014: combine physical time with a logical counter to get timestamps that are close to real time but still safe under clock skew. You get the best of both worlds: causality when you need it, and real-time affinity. muratbuffalo.blogspot.com/2014/07/hybr...

Hybrid Logical Clocks Here I will write about our recent work on Hybrid Logical Clocks, which provides a feasible alternative to Google's TrueTime. A brief hist...

I built a small interactive visualizer for Hybrid Logical Clocks (HLC).

I used Claude Code to put this together quickly and make the behavior visible step by step.

Try it here:

muratdem.github.io/hlc-visualiz...

Feedback welcome. Share if useful.

Measuring Agents in Production When you are in TPOT echo chamber, you would think fully autonomous AI agents are running the world. But t his 2025 December paper, "Measur...

[new blog post]

Measuring Agents in Production

muratbuffalo.blogspot.com/2026/03/meas...

Modeling Token Buckets in PlusCal and TLA+ Retry storms are infamous in distributed systems. It is easy to run into them. Inevitably, a downstream service experiences a hiccup, so you...

[new blog post]

Modeling Token Buckets in PlusCal and TLA+

muratbuffalo.blogspot.com/2026/03/mode...

LOL... My version is competent and organized. Yours has a voice.

Frank if you would be rendered obsolete, I wonder who wouldn't be??

The Serial Safety Net: Efficient Concurrency Control on Modern Hardware This paper proposes a way to get serializability without completely destroying your system's performance. I quite like the paper, as it fli...

[new blog post]

The Serial Safety Net: Efficient Concurrency Control on Modern Hardware

muratbuffalo.blogspot.com/2026/03/the-...

This paper proposes a way to get serializability without completely destroying your system's performance. It flips the script on how we think about isolation levels.

What Does Privacy Mean Now? Do you care if an LLM sees you naked?

Do you care if an LLM sees you naked? What does privacy mean now? emptysqua.re/blog/what-do...

Now TLA+ is on VS Code.
Just load the TLA+ plugin on VS Code.
It is a much better experience.

TLA+ as a Design Accelerator: Lessons from the Industry After 15+ years of using TLA+, I now think of it is a design accelerator. One of the purest intellectual pleasures is finding a way to simpl...

[new blog post]

TLA+ as a Design Accelerator: Lessons from the Industry

muratbuffalo.blogspot.com/2026/03/tla-...

TLA+ Tilapias
#tlaplus

Monster Scale Summit Agenda Agenda Session times are displayed in your local timezone. Agenda subject to change. Wednesday 3/11 Thursday 3/12 Instant Access

Monster Scale Summit (free and virtual) starts tomorrow. Pop in to learn from an incredible lineup (@skamille.themanagerswrath.com, @antirez.bsky.social, @martin.kleppmann.com, @muratdemirbas.bsky.social, @dominiktornow.bsky.social, @teivah.dev Pat Helland, Joran Greef....) www.scylladb.com/summit

Building a Database on S3 Hold your horses. I'm not unveiling a new S3-native database. The paper is from 2008. Many of its protocols feel clunky today. Yet it nails ...

[new blog post]

Building a database on S3

muratbuffalo.blogspot.com/2026/03/buil...

a screenshot of an excerpt from an interview with Murat Demirbas about writing engineering blog posts

New tech blogger interview -- this time we’re featuring @muratdemirbas.bsky.social, who’s about to publish his 800th (!) blog post
writethatblog.substack.com/p/murat-demi...

800th blog post: Write that Blog! I had given an email interview to the "Write That Blog!" newsletter. That came out today , which coincided with my 800th blog post. I am inc...

[new blog post]

800th blog post: Write that Blog!

muratbuffalo.blogspot.com/2026/03/800t...

[new blog post]

Writing Away From the Screen

muratbuffalo.blogspot.com/2026/02/writ...