For "Consistent hashing with bounded loads", how does a READ operation find which is the fallback node for a particular key?
I assume it cannot directly use the output of the hash function since that was the overloaded node.
I see 2 problems regarding time-varying "data to node" assignments.
One is the overloaded node is no longer overloaded, so fallback is not needed anymore. Second is even the fallback is overloaded so we fallback again.

Maybe we can associate the location of the fallback node with the key after writing to a node?

The nodes are assigned to multiple positions on the hash ring by hashing the node IDs through distinct hash functions to ensure uniform distribution of keys among the nodes

Rather than using different hash functions to hash the same server node ID (key), I believe it would be much more optimal to use the same hash function but update the server node Id (key) by appending some distinct values.

For example:
Server-node-1
Server-node-1#A
Server-node-1#B
.....Server-node-1#AA etc.

Adding a node:

Step 2: Identify the keys that fall within the subrange of the new node from the successor node on BST

When comparing the keys, the algorithm typically performs a search operation in the BST to identify only the keys that fall within the subrange of the new node. This search process takes advantage of the BST's properties, such as logarithmic time complexity for search operations, to quickly locate the relevant keys for transfer between the nodes.

By utilizing the inherent structure of the BST and its efficient search capabilities, the comparison process can accurately pinpoint the subset of keys that need to be moved without having to individually compare each and every key in the successor node's BST against the new node's key range. This optimized approach helps streamline the identification and transfer of keys during the insertion process on the hash ring.

wow! I had read consistent hashing from many resources earlier. No one covered it in depth. so I read the white papers to understand the details. You have covered all the details. I wish this was available in 2022! Awesome Neo!

Thanks!!