Can a computer system be perfectly consistent, available, and partition-tolerant all at once?
Image: photo taken by: de:User:Stahlkocher, CC BY-SA 3.0, via Wikimedia Commons
Krull–Akizuki theorem
Can a computer system be perfectly consistent, available, and partition-tolerant all at once?
Imagine you're streaming a movie online. You want it to always play smoothly (consistency), even if your internet goes down (partition tolerance), and never miss a scene (availability).
The CAP theorem explains that in a distributed system, you can't have all three: consistent, available, and partition-tolerant at the same time. It's like trying to balance a seesaw with three weights on it; you can only have two weights balanced at once.
Example
If you prioritize consistency and partition tolerance, you might have to deal with buffering (lack of availability) when there's a network issue.
Remember this
You can only optimize two out of the three: consistency, availability, or partition tolerance.
Text adapted from Wikipedia, licensed under CC BY-SA 4.0.
Related concepts
CAP theorem
Can you always access your favorite online photo album from anywhere?
CAP theorem states: you can have at most 2 of consistency, availability, partition tolerance
Ever wondered why you can't always get the latest news instantly?
the blockchain trilemma says
The blockchain trilemma posits you can optimize only 2 of decentralization, security, and scalability
Cronbach's alpha
Cronbach's alpha (α) measures internal consistency
Coase theorem
Can strangers fix a broken window without calling the cops?
Turing machine
Can a machine think like us?
Educational content, not financial advice.
Swipe through 100 ML concepts daily
Open Pocket Polymath