difference between exactly-once and at-least-once guarantees

Question

I'm studying distributed systems and referring to this old question: stackoverflow link

I really can't understand the difference between exactly-once, at-least-once and at-most-once guarantees, I read these concepts in Kafka, Flink and Storm and Cassandra also. For instance someone says that Flink is better because has exactly-once guarantees while Storm has only at-least-once.

I understand that exactly-once mode is better for latency but at the same time it's worse for fault tolerance right? How can recover a stream if I haven't duplicates? and then... if this is a real problem, why exactly-once guarantee is considered better than others?

Someone can give me better definitions?

user_1410501 · Answer

I've found a great website where all (or most) Cloud Computing Patterns are succinctly discussed. I really recommend it to you, take a look: http://www.cloudcomputingpatterns.org

Exactly-once Delivery

For many critical systems duplicate messages are inacceptable. The messaging system ensures that each message is delivered exactly once by filtering possible message duplicates automatically.

At-least-once Delivery

In case of failures that lead to message loss or take too long to recover from, messages are retransmitted to assure they are delivered at least once.

difference between exactly-once and at-least-once guarantees

4 Answers