This article is a serious example of the difficulties encountered in using MongoDB, and the reasons for eventually discarding MongoDB.
Spent a lot of time to read, but have not carefully read the review, and so on which day is not busy, you can analyze the information inside.
Here is a brief list of the author's views:
1. MongoDB's document is suitable for storing self-contained data, does not apply to the scene of a join between document, in this scenario, the RDBMS + cache is the most effective, there is no redundant data, but also through the Background app automatically update the cache to improve the query performance of the program, Avoids query inefficiencies caused by multi-table joins
2. The faster way to query is to put the active stream data into a document to find, MongoDB is a kind of, the previous cache is similar method
3. Social data is a graph in nature, but the relational type can still be stored efficiently, so the RDBMS + cache approach is also good
When MongoDB should not be used