The document discusses different technologies for storing and querying large chemical datasets, known as "big chemical data". It evaluates PostgreSQL, SQLite, MessagePack, FlatBuffers, and Pandas on a test dataset of 4 million compounds from ZINC. For queries like retrieving atom counts for 50k molecules, counting molecules by atom number, and fingerprint lookups, SQLite and MessagePack performed the fastest, completing in under 50ms. PostgreSQL was also very fast with indices, finishing some queries in under 100ms. The document concludes no single technology is best and the complexity of the tool should match the task.