Massive Data Exchange Between Bounded Context (Microservices)

[Fabio Corvino]

Hi,

In a microservices architecture, we are implementing a microservice for scheduling orders. Naturally, each microservice has its own database, and the scheduling one contains only data related to scheduling parameters.

The Scheduling microservice does not operate in the background, but is used by some users who manually interact with and alter the algorithm's results.

The challenge we face is that the scheduling microservice needs to access data belonging to two other microservices. Normally, I would opt for reading via API from the respective services, as it functions like a 'read model'. However, in this case, the data are extensive, and users will experience a significant delay when loading data for a new plan.

Some colleagues suggest creating a copy of the data in the scheduler's database. This data would be kept updated with all the update events from the other microservices. However, this approach would generate a considerable amount of traffic (especially when data are synchronized from the ERP), which is not essential for the scheduler domain. The user works with stale (in-memory) data and decides when to update the plan he is working on. Furthermore, this pattern requires maintaining all interfaces for the update events.

I am leaning more towards using a NoSQL cache database, like a Redis cache, in the calling service, which we can refresh occasionally. This cache would be populated by calling the APIs of the other microservices.

Do you have any suggestions?

Many thanks!

Fabio