Course contentsShow
← System Design
Lesson 409 of 1,91911. NoSQL: Document StoresFree lesson

Data Size and Storage Considerations

How JSON overhead and document bloat affect storage costs and performance at scale.

Data Size and Storage Considerations

You'll learn how JSON overhead and document bloat affect storage costs and performance at scale in document stores.

The Hidden Cost of Flexibility

Document stores use formats like JSON or BSON that embed field names inside every single document. Unlike relational databases where column names are stored once in a schema, document stores repeat field names in each record. This creates storage overhead.

Imagine storing a million user profiles. In a relational table, you define columns like firstName, lastName, and email once. In a document store, these field names are written a million times:

{"firstName": "Alice", "lastName": "Smith", "email": "alice@example.com"}
{"firstName": "Bob", "lastName": "Jones", "email": "bob@example.com"}
// ... repeated a million times

Each field name consumes bytes. Long, descriptive field names (like userPreferencesSettings) multiply this overhead.

Document Bloat

Embedding documents (as you've learned) adds convenience but increases size. A product document with embedded reviews, images, and metadata can balloon to megabytes. Large documents:

  • Consume more disk space and memory
  • Take longer to read and write
  • Slow down network transfers between shards and replicas
  • Increase replication and backup costs

Storage Implications at Scale

At small scale, overhead is negligible. At billions of documents, it translates to terabytes of extra storage. This affects:

  • Infrastructure costs: More disks, more cloud storage fees
  • Query performance: Reading large documents from disk takes time
  • Working set size: Fewer documents fit in RAM, reducing cache effectiveness

Mitigation Strategies

Use shorter field names for high-volume collections. Avoid unnecessary embedding. Compress data where possible. Monitor document sizes during development.

Key Takeaway: Document store flexibility comes with JSON overhead and potential bloat—repeated field names and embedded data increase storage costs and can degrade performance at scale, requiring careful schema design and monitoring.