This pattern of cluster is created for applications where reads are very frequent and writes are less frequent. Another condition necessary for this pattern to work is to have a limited size of data, or in other words, data capacity that can fit into the hardware provisioned for the master (the same hardware configuration is needed for the slaves too). Since the requirement is to cater to frequent reads, this pattern also has the capability to scale horizontally. Another point we have to keep in mind is that replication in slaves can have a time delay which can result in stale data getting served. The business requirement should be okay with that scenario.

The solution for this pattern would be to have all the writes done to the master, and have the slave cater to all the reads. The reads to slaves need to be load balanced so that performance can be met.

Redis provides an in-built capability to have master-slave configuration wherein the writes can be done to...