BASE Model

NoSQL (actually short for “not only” SQL) often times rely on a softer transaction model than ACID abbreviated as BASE:

  • Basically Available
  • Soft State
  • Eventually Consistent

Basically Available

This has to do with the perceived availability of the data. The idea is for data to always be available even if it is not in a consistent or hardened state. There are no isolation levels. One transaction may access a file that is being actively modified by another.

Soft State

Values stored in the database may change because of the eventual consistency model.

Eventually Consistent

Data and changes to said data are replicated across nodes and eventually become consistent. For example, in Hadoop, when a file is written to the HDFS, the replicas of the data blocks are created in different data nodes after the original data blocks have been written. For the short period before the blocks are replicated, the state of the file system isn’t consistent.

As tongue in cheek as BASE is, it’s a bit misleading. It is not meant to be the polar opposite of ACID because most NoSQL implementations do not abandon all ACID principles. Soft State and Eventually Consistent are also so closely related that their distinction serves no real purpose.

The point though, is that a BASE model uses a more relaxed approach to consistency for the sake of being able to scale horizontally much more easily and provide a high degree of availability.

This model is very popular with social media platforms and search engines where “some (maybe partial) results now” is much more important than “all results no matter how long it takes”.

CAP Theorem

Also known as Brewer’s theorem, named after computer scientist Eric Brewer, proposes that it is only possible for a distributed data store to provide two out of the three following guarantees:

Consistency: Every read receives the most recent write or an error

Availability: Every request receives a (non-error) response, without the guarantee that it contains the most recent write

Partition tolerance: The system continues to operate despite an arbitrary number of messages being dropped (or delayed) by the network between nodes