← All series

Series

Databases from the Engineer's Perspective

Storage engines, indexes, query lifecycle, transactions, and concurrency — from an application engineer's perspective, grounded in production experience with both PostgreSQL and MongoDB.

4 published · 1 upcoming

  1. Part 1Published

    Storage Engines: How PostgreSQL and MongoDB Actually Write to Disk

    Most engineers treat the database as a black box — until production breaks. This article builds a deep mental model of how PostgreSQL (heap + WAL + MVCC) and MongoDB (WiredTiger B-tree + journal) actually store and retrieve data, and how those design choices shape every architecture decision you make.

  2. Part 2Published

    How Indexes Work Under the Hood: B-Tree, Hash, and Beyond

    Indexes are the single most misunderstood tool in a database engineer's toolkit. This article goes deep into how B-Tree, Hash, GIN, and partial indexes physically work in both PostgreSQL and MongoDB — and why that knowledge changes the way you design schemas, write queries, and diagnose performance regressions.

  3. Part 3Published

    The Query Lifecycle: From Your Application Code to Disk and Back

    Every query your application sends travels through a precise sequence of stages before results come back. Understanding that sequence — parsing, rewriting, planning, execution, and result transport — is what separates engineers who guess at performance problems from engineers who diagnose them systematically.

  4. Part 4Published

    Transactions, ACID, and What Consistency Actually Costs You

    ACID is one of the most cited and least understood guarantees in database engineering. This article goes beneath the acronym — into what atomicity, consistency, isolation, and durability actually mean at the storage level, what each isolation level permits and prevents, and how PostgreSQL and MongoDB make fundamentally different trade-offs on the same concurrency problems.

  5. Part 5In progress

    Concurrency Control: How PostgreSQL and MongoDB Handle Simultaneous Writers

    Concurrency control is where database theory meets production reality. This article goes deep into how PostgreSQL and MongoDB handle simultaneous readers and writers — the full MVCC mechanism, every lock type, how contention manifests, and the systematic approach to diagnosing and resolving it before it becomes an incident.