I have a very large geospatial index in redis that I replicate to application servers and query using SINTERSTORE operations on read-write replicas.
I would like to understand if redis server-assisted client-side caching could be used as an effective substitute for this architecture.
Specifically, if I use the BCAST option and specify no prefix, have I effectively achieved the same thing as redis replication? Based on the article above,
If no prefix is specified at all, the prefix is assumed to be the empty string, so the client will receive invalidation messages for every key that gets modified.
Every time a key matching any of the prefixes is modified, all the clients subscribed to such prefix, will receive the invalidation message.
I would note that replication has extensive discussion about what happens where there are network disruptions; no such conversation exists WRT server-assisted client-side caching. I wonder if this is one of the differences?