A useful reminder on the CAP Theorem (aka Brewer theorem), as I was studying a MOOC on HDFS.
What the CAP Theorem says is that on a distributed system, it is impossible to guarantee more than 2 out of the three following constraints:
- Consistency: every read receives the most recent version of the data
- Availability: every read receives an answer, without guarantee that it contains the most recent write.
- Partition tolerance: the system operates even when the network between the nodes fails(except in case of a total breakdown).
The following DBMS can be positioned as follows:
- C+A : Oracle, MySql
- C+P : HDFS, MongoDB
- A+P: Redis, Elasticsearch
