Rendezvous also known as highest random weight (HRW) hashing implementation.
It's a non so well known alternative to well-known consistent hashing(even though rendezvous hashing precedes consistent for around a year).
As a consistent hashing algorithm it maintains K/n remapping invariant(K/n keys are remapped whenever the list of endpoints changes (K is the total number of keys and n is the total number of nodes/servers)).
Not really, the main problem with rendezvous hashing is that it's somehow hard to deal with the hotspot problem, should it arise.
- Lightweight memory consumption, the only thing that you need to store in a non-optimized version is the list of nodes.
- It's easier to understand and implement.
- Provides a good distribution without a lot of complexity(it also depends on the hash functioin that you use to hash (node,key) pair).