-1

Whats difference between fail-over and replication? I tried reading from this article https://github.com/donnemartin/system-design-primer#availability-in-numbers but I could not understand the difference.

adwaita jadhav
  • 1

1 Answers1

0

Replication is creating or maintaining multiple copies of something -- generally your database, but possibly more, such an an image of your entire server.

Failover is when one system detects that another has failed, and responds by taking over its duties.

They are completely different things, though they are often are used together to serve purposes such as fault tolerance and disaster preparedness.

John Bollinger
  • 160,171
  • 8
  • 81
  • 157
  • 1
    So, it's like keeping a spare tyre in a jeep with air-filled (i.e. latest data copied)... which is called `replication`. And, when one of the running tyres gets punctured/burst, you replace it with this spare one, which is a kind of `fail-over` mechanism. Now, in the case of tyres, it needs manual intervention, for software, it's mostly automated. – ms_27 Sep 28 '21 at 12:52
  • 1
    Yes, @ms_27. But it can also be like a large truck or trailer that has four tires on an axle instead of two. That's a replication of one tire on each side. Under normal conditions, they all support the load, but if one tire fails then the other three can bear the resulting extra load until the failed one is replaced. Both approaches are used, and they can even be used together. – John Bollinger Sep 28 '21 at 13:04
  • Agreed. This is a more relevant real-life example for automatic fail-over. – ms_27 Sep 28 '21 at 13:42