3 ms·
The predecessor to this was running one hash function per cluster member and picking the one with the highest or lowest result. Personally I still find that a
by hinkley 2y ago
The predecessor to this was running one hash function per cluster member and picking the one with the highest or lowest result.
Personally I still find that a lot easier to reason about. Especially when it’s time to resize the cluster.