The Eventually Consistent Paradigm

History

...computer scientist Eric Brewer (UC Berkeley), the theorem first appeared in fall 1998. It was published as CAP principle in 1999 and presented as a conjecture by Eric Brewer at the 2000 Symposium on Principles of Distributed Computing (PODC). In 2002, Seth Gilbert and Nancy Lynch of MIT published a formal proof of Brewer's conjecture, rendering it a theorem. Wikipedia

It is impossible for a distributed computing system to simultaneously provide:

• Consistency
• Availability
• Partition tolerance

Facebook, Twitter, et. al. scale by addressing CAP limitations in various ways.

It isn't always by relaxing consistency

Consistency

A service that is consistent operates fully or not at all. (see Julian Browne)

Example:

• ...when you post your tweet,
• everyone has the same view,
• your view of your tweet history is the same view everyone else has of it as well.

Availability

Your service is always available.

Example:

• You login to Twitter,
• you post your tweet,
• you see success,
• you move on.

Partition tolerance

No set of failures less than total network failure is allowed to cause the system to respond incorrectly (Gilbert & Lynch)

Examples:

• excessive latency or severed network connectivity,
• separate nodes in your infrastructure,
• yet your system continues to function normally

Interpretations Vary

Recommended read: CAP Twelve Years Later: How the "Rules" Have Changed

The "2 of 3" formulation was always misleading because it tended to oversimplify the tensions among properties. Now such nuances matter. CAP prohibits only a tiny part of the design space: perfect availability and consistency in the presence of partitions, which are rare.

The bigger your data, the more you need to worry about the CAP Theorem.

Yet, we've come a long way.

ACID / BASE are more local properties than emergent system properties.