How to use this reference
- Work through topics top to bottom — storage internals and MVCC assume you understand relational schema shape; sharding assumes you understand replication.
- DDIA and Database Internals are anchor books — each subtopic points to specific chapters, not cover-to-cover reading.
- We deliberately do not repeat the SQL query-writing track, generic distributed-systems theory (Raft, sagas), or OS file-system deep dives — those are separate prerequisites or companions.
- Scope is storage only — no Spark pipelines, Airflow orchestration, or stream processing. Data lakes and table formats are included because they define how files are organized and versioned on disk.
The Reference
- 1
How relational data is shaped before it ever hits a query planner — normalization trade-offs, keys, and index design from a storage perspective.
- 2
What happens below the SQL layer — B-tree vs LSM engines, write-ahead logs, pages, and the buffer pool.
- 3
ACID guarantees, isolation levels, and how Postgres-style MVCC actually stores row versions.
- 4
How RDBMS products replicate data and fail over — the database lens on a problem whose theory lives in Distributed Systems.
- 5
When one Postgres isn't enough — horizontal partitioning, shard keys, and how Vitess and Citus scale relational storage.
- 6
Document and wide-column stores — when relational rows aren't the right storage primitive.
- 7
Cache-aside through write-behind, stampede mitigation, and Redis, DynamoDB, graph, and time-series engines.
- 8
S3, HDFS, Parquet, and table formats — how analytical data is stored at petabyte scale without a traditional database server.
- 9
Inverted indexes for full-text search and columnar engines for analytics — two storage layouts optimized for very different read patterns.