The simulator demonstrates how Rendezvous Hashing assigns each key to a node by computing an independent pseudo-random weight per node and selecting the maximum, and how adding or removing a node only reshuffles the specific keys that node was involved in winning, leaving all other key assignments completely stable.
Add or remove nodes using the node controls, then insert keys one at a time or in a batch to see each node's computed weight for that key with the highest value highlighted as the winner. Remove the currently winning node for a key and watch the reassignment happen live, or add a new node and observe which existing keys it captures based purely on its own computed scores.
Add node, remove node, insert single key, insert batch of keys, toggle score display per node, and reset the simulation to an empty node set.
Because Rendezvous Hashing needs no shared ring or routing table, any client that independently knows the current node list will always compute the exact same key assignment as every other client, which is why the technique has been used in coordination-free request routing at several large-scale caching and content delivery systems.
The simulator demonstrates how Rendezvous Hashing assigns each key to a node by computing an independent pseudo-random weight per node and selecting the maximum, and how adding or removing a node only reshuffles the specific keys that node was involved in winning, leaving all other key assignments completely stable.
The simulator demonstrates how Rendezvous Hashing assigns each key to a node by computing an independent pseudo-random weight per node and selecting the maximum, and how adding or removing a node only reshuffles the specific keys that node was involved in winning, leaving all other key assignments completely stable.
Add or remove nodes using the node controls, then insert keys one at a time or in a batch to see each node's computed weight for that key with the highest value highlighted as the winner. Remove the currently winning node for a key and watch the reassignment happen live, or add a new node and observe which existing keys it captures based purely on its own computed scores.
Because Rendezvous Hashing needs no shared ring or routing table, any client that independently knows the current node list will always compute the exact same key assignment as every other client, which is why the technique has been used in coordination-free request routing at several large-scale caching and content delivery systems.