RendezvousNet

Rendezvous also known as highest random weight (HRW) hashing implementation.

What is rendezvous hashing?

It's a non so well known alternative to well-known consistent hashing(even though rendezvous hashing precedes consistent for around a year).

What is it good for?

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)).

Is it better than consistent hashing?

Not really, the main problem with rendezvous hashing is that it's somehow hard to deal with the hotspot problem, should it arise.

Why use rendezvous hashing then?

1. Lightweight memory consumption, the only thing that you need to store in a non-optimized version is the list of nodes.
2. It's easier to understand and implement.
3. Provides a good distribution without a lot of complexity(it also depends on the hash functioin that you use to hash (node,key) pair).