MongoDB Tips

  1. About data model design [1][http://docs.mongodb.org/manual/core/data-model-design/]

There are two data model: embedded data models and normalized data models. We compare them in performance, space, model relationship, atomicity.

* Performance: Embedded data models (E) is faster than Normalized
  data models (N) because no matter reading, updating, the former 
  just need 1 ops. N need 2 or more.

* Space: E need more disk/memory space than N.

* Model relationship: E applies 1-to-1 or 1-to-many relationships.
  while N applies many-to-many relationships.

* Atomicity: Because updating one document is atomic, E need just 1
  op to update its data, so E ensure its atomicity. While N need more
  than 1 op to update data in separate documents, we can't ensure
  more 2 ops' atomicity.
Published at 05 March 2014