PACELC Theorem

Last Updated : 21 Sep, 2022

The PACELC Theorem is an extension of the CAP Theorem. One of the questions that CAP Theorem wasn’t able to answer was “what happens when there is no Partition, What Logical Combination then a Distributed System have?“. So to answer this, In addition to Consistency, Availability, and Partition Tolerance it also includes Latency as one of the desired properties of a Distributed System. The acronym PACELC stands for Partitioned, Availability, Consistency Else Latency, Consistency.

PACELC Theorem:

PACELC theorem states that in the case of Network Partition ‘P’ a distributed system can have tradeoffs between Availability ‘A’ and Consistency ‘C’ Else ‘E’ if there is no Network Partition then a distributed system can have tradeoffs between Latency ‘L’ and Consistency ‘C’.

As we can clearly see from the diagram the ‘Yes ‘ part is the CAP Theorem and the ‘No’ part is the extension to the CAP Theorem which is PACELC Theorem. For more details on CAP theorem you can refer Various properties of CAP Theorem article.

Partition:

Partition basically means two 2 nodes are not able to communicate with each other. The below diagram gives a better understanding of the same.

Latency in PACELC Theorem:

One of the major pitfalls of the CAP Theorem was it did not make any provision for Performance or Latency, in other words, CAP Theorem didn’t provide tradeoffs when the system is under normal functioning or non-partitioned. Let’s try to understand this by an example:

Consider a scenario, in which you are making a request to Signup up and getting the confirmation after an hour. According to the CAP theorem, this system is available but such latency is unacceptable in any real-world application.

Advantages:

PACELC Theorem considers the latency and consistency tradeoffs that were always prevalent.
It helps us to think about the effective way to choose, and design distributed systems.
Overcomes the major pitfalls of the CAP Theorem.

Conclusion:

All in all PACELC Theorem gives a designer two modes of operation, with and without partitioning. With partitioning, one must choose between Consistency and Availability otherwise in the case of normal operation one must choose between Latency and Consistency.