The original title is "Inside LioranDB: Why the Storage Engine Speaks Bytes, Not JSON"

Short summary

LioranDB's storage engine operates purely on table, key bytes, and value bytes, with no understanding of documents or JSON. The DBMS layer above adds document-oriented features like collections, queries, secondary indexes, and transactions on top of the transactional key-value engine. This separation keeps the core engine small, reusable, and focused on low-level concerns like durability, page management, and recovery.

• •Storage engine handles only table + key bytes + value bytes
• •DBMS layer adds document features: collections, JSON, queries, indexes
• •Layered separation keeps engine small and reusable for future data models

Generated with AI, which can make mistakes.