0

I have a question. I read many blogs, websites about CAP theorem. They say that single-node systems are CA, but how can it be "A" if that single node goes down ? Because if it does, the system will be unavailable, right ? And is "availability" in CAP equal to "availability" in HA of distributed systems ?

Sake
  • 96
  • 5
  • i think this is covered by partition-*in*tolerance. partition tolerance basically means that the system is operational if parts of the system go down. if a single node system goes down, and it is not operational, then it is not partition tolerant. – Effie Dec 27 '21 at 15:00
  • also, i think that this question should be moved to: [https://cs.stackexchange.com/](https://cs.stackexchange.com/) – Effie Dec 27 '21 at 15:01
  • @Effie I have understood why it's not "P", but how about "A" ? if the single node system goes down, it will violate "A" ? – Sake Dec 27 '21 at 16:11
  • i guess it is consistent and available at all times at which there is no partition. and there is no guarantees about anything when there is a partition. – Effie Dec 27 '21 at 16:24
  • but you should move the question, i think there are people there who know better :) – Effie Dec 27 '21 at 16:36
  • @Effie I will try it. Thanks – Sake Dec 27 '21 at 17:25

1 Answers1

0

That is not the definition of (cap) availability, it is not about being able to reply but refusing to do so, e.g. the node receives a message and it is not able to contact the rest of the cluster, what can it do: i) respond thereby becoming A but not C; ii) do not respond and become C but not A. See my previous answer here for illustrative examples.

As for your specific example I'm not sure it makes sense to reason about CAP if the cluster does not exist, I believe that you need at least one node.

João Matos
  • 6,102
  • 5
  • 41
  • 76