What Is CAP Theorem?

The CAP theorem sounds severe, but it's pretty simple. It states that no matter how hard you try in a distributed computer system, you can only have three guarantees: consistency, availability or partition tolerance. The CAP theorem states that a system must choose between these three properties as it cannot have all three simultaneously. In distributed systems, the loss of consistency is usually more critical than the loss of availability or partition tolerance. The theorem, also known as Brewer's theorem, basically says you can't have a distributed computer system that simultaneously provides all three of consistency, availability and partition tolerance at optimal levels. Significant efforts are put into designing consistent systems. Therefore, depending on the requirements, developers must choose the right combination of these three traits for their strategies. It is an idea outlining the limits of distributed computing. It states that one must sacrifice either character or availability to have less than 100% network partition tolerance. As networks become more extensive (or other criteria more stringent), this must increasingly become so. It is the heart of computer science. It's not just a set of letters like ABC or DEF. CAP covers the rules for ensuring that your data is consistent and up to date, no matter which node currently hosts it. If you have a copy of a piece of data on your computer, are you sure everyone else has the same version? Paraphrasing CAP Theorem expert Terry Halpin: If you can't guarantee C (Consistency) or A (Availability), then your database is providing P (Partial) availability or consistency when measured against its stated intent. There are, however, other ways of thinking about the CAP theorem. One way is that it's a bit like the old tradespeople's motto: you can have services cheap, fast and of good quality - but not all three. The CAP theorem points to the same idea - limited resources and results.