One of the ways to help microservices scale is to outsource some of their tasks. One such task that may hinder scaling efforts is memory management and caching data.
For a single monolithic application, storing cached data directly in the process memory is not a problem as the process will be the only one accessing the cache anyway. But with several replicas of a process, this approach starts to show some problems.
What if one replica has already computed a piece of a workload and stored it in a local cache? The other replica is unaware of this fact and has to compute it again. This way, your application wastes both computational time (as the same task has to be performed multiple times) and memory (as the results are also stored with each replica separately).
To mitigate such challenges, consider switching to an external in-memory store rather than managing the cache internally within an application. Another benefit of using an external solution is that...