Ayder: A Durability-First Event Log System Verified for Crash Recovery
By
Aydarbek
4mo ago· 4 min readenCode
95/100
Golden Brown
Bagelometer↗
Pure flour-power. Hearty enough to carry you through lunch.
Score95Typepress releaseSentimentpositive
Summary
Ayder is a durability-first event log system designed to maintain data correctness under real-world failures including crashes, partitions, and disk failures. Unlike systems optimized for throughput, Ayder prioritizes verified durability, having passed 45/45 mixed-fault Jepsen tests. It's a single binary, HTTP-native, Raft-backed system that ensures acknowledged data remains durable even after process crashes mid-write. The system addresses common failure handling uncertainties in edge and IIoT environments and includes a live crash demo for testing.
Key quotes
· 4 pulledAyder is built for something else: correctness under crashes, partitions, and disk failures.
Verified with Jepsen: 45/45 mixed-fault tests passed
SIGKILL → restart → data still correct
what happens if the process crashes mid-write? is acknowledged data actually durable? how is recovery verified?
Contribute to A1darbek/ayder development by creating an account on GitHub.
