This chapter follows MongoDB cheat sheet. It closes the MongoDB series.
BSON
Binary JSON: MongoDB’s on-disk and on-wire document encoding. Supports types JSON lacks (ObjectId, Date, Decimal128). See CRUD operations.
ObjectId
Default _id type: 12 bytes, unique per shard, roughly time-ordered. See CRUD operations.
Collection and database
Named container for documents; databases group collections. Created implicitly on first write. See CRUD operations.
WiredTiger
Default storage engine: document-level concurrency, compression, checkpoints. See A short history of MongoDB.
Replica set
Group of mongod nodes with one primary and zero or more secondaries for redundancy and failover. See Replica sets.
Oplog
Capped collection on each replica set member recording operations for replication. See Replica sets.
Write concern
Acknowledgment rules (w: 1, majority, etc.) controlling durability visibility. See Replica sets.
Shard
Horizontal partition of data in a sharded cluster; routed by shard key. Mentioned in history; deep sharding ops are outside this series.
GridFS
Specification for storing files larger than 16 MB split across fs.files and fs.chunks collections.
Aggregation pipeline
Sequence of stages transforming documents ($match, $group, …). See Aggregation Framework.
Search index
Atlas Search index definition (analyzers, field mappings) powering $search. See Full-text search.
Embedding
Numeric vector representing semantic content of text, used with $vectorSearch. See Vector search and embeddings.
SCRAM
Password authentication mechanism used by MongoDB users. See Authentication.
Session
Driver handle grouping operations in a transaction with snapshot isolation. See Transactions.