Engineer by craft, explorer by instinct

Blooooooog

Thoughts, deep-dives, and learnings from building software.

Strong Consistency Usually Kills Read Scalability — CRAQ Challenges That

How CRAQ builds on chain replication to scale reads across replicas without giving up strong consistency, and why this idea is still one of the cleverest designs in distributed storage systems.

Apr 23, 2026

Distributed SystemsStorage SystemsConsistency

Why Adapter Is My Favorite Design Pattern

How the adapter pattern quietly solves real-world integration problems—payments, third-party APIs, and libraries by isolating change and keeping your core system clean.

Mar 18, 2026

Design PatternsSystem DesignSoftware Engineering

ZooKeeper Under Contention — Correctness Is Easy, Throughput Isn’t

What actually happens when hundreds of clients try to update the same znode, why retries explode, and how different concurrency strategies behave in a real ZooKeeper cluster.

Feb 24, 2026

Distributed SystemsZooKeeperConcurrency

The Hardest Part of Raft Is Not Elections — It’s Log Replication

Why most Raft implementations don’t fail during leader election, but break down while replicating logs—when nodes diverge, leaders crash mid-flight, and consistency has to be actively repaired under real-world conditions where nothing stays perfectly in sync.

Jan 9, 2026

Distributed SystemsRaftConsensus