This week, I presented this paper at ICCCN'26. This is joint work with Aleksey Charapko (University of New Hampshire) and my MongoDB colleagues Matt Broadstone, Daniel Gomez Ferro, and Akshat Vig. The paper investigates how to build a metastability tolerant execution control system (ECS) for a...
The Aurora DSQL paper finally dropped. Reading it yesterday was an interesting experience, because I spent two years (2022-23) working with the AWS team that designed and built Aurora DSQL. Since I have been very familiar with the architecture, the paper's overview description of the system didn't...
As I mentioned in my previous post, I traveled to San Jose at the end of May for the ACM CAIS conference. On Day 0, I gave a very short talk at the Supporting our AI Overlords (SAO) workshop. This post is the promised summary of our paper, "A Case for Simulation-Driven Resilience in Agentic Data...
Introduction If you've ever dealt with ANY database system you probably encountered the ACID acronym. Let's briefly go over each of the principles it describes: A for Atomicity The concept of treating something (a block of data operations in our context) as a single, indivisible unit of work. C...
Update (2026): the thinking in this post is now carried forward in Minigraf Getting Started with Time Time is of some importance to humans. The best scientific minds have yet to unravel all its mysteries. Even its definition is a recurring subject of debate. Yet, we all innately have a sense of...
“Can we do to web services what Linux did to the operating systems?” Imagine what the web with a P2P GitHub, a P2P YouTube, a P2P Twitter, a P2P Google search, and so on. But getting started with P2P services remains difficult, because: “There is no P2P PostGRES!” This is a very hard problem, so...
Multi-version histories In conventional histories, at any point in time, there exists only one version of a data item. That means that the write operations always replace current version and the read operations always read the latest written version as it is defined by the schedule semantics:...
Once upon a time, I was interviewed for a database engineering position. I was asked about what deadlock prevention algorithms do I know. I could not name a few (besides detection and randomization) and I was rejected. It was an unsatisfying experience because the ideas behind those algorithms...
Recently, I came across an excellent exercise of testing a scheduling history for belonging to different serializability classes and had a lot of fun drawing a polygraph and finding out a cycle in it. When I finally managed to draw the correct version of the polygraph, I thought that it could be...
This document contains simple examples and explanations of transaction isolation anomalies well known since the late ’90s. The work based on papers [1], [2], [3]. I hope it will be helpful for someone as a quick reference. List of covered anomalies include: G0: Dirty writes G1: Dirty...
I favor strict national gun regulation. There are bad ways to do it, and good ways, but almost any consistent regime would be an improvement over the status quo. If you had asked me the day before yesterday, I would have said that I'd support nearly any gun control measure that was brought before...