Performance improvement of redis transactions


We are looking for ways to improve the performance of redis transactions. The intention is to look at specific location for a particular data set instead of looking at all the entries in the server. Mostly data that we store is in the form of key value pairs. Looking for optimization options like hashing, partitioning and indexing to improve the performance of the transactions.

We are using a clustered aws elaticache setup; also using java 11 and lettuce framework to read/write data to redis.