In Partition based on the hash of the term approach, let’s say 2 requests for CAP and CAPITOL are coming, as the hashed value of both the terms will be different so the request can go to 2 different servers. This means trie starting with CAP needs to be stored in both the servers.
Basically, we will need to store complete trie in all the servers, and search throughput will be equally divided based on hash key.
Pleasse correct me if I am wrong.