Modeling data for operational performance

So far, we have focused on breaking up the monolith and materializing data downstream. Now, we will zero in on modeling data to support the access patterns needed during the operational phases of the data life cycle. This is where various actors (that is, users) create and use data via BFF services. These early phases of the data life cycle need particular attention because they have seen significant changes and improvements.

First, we will look at an approach to logical data modeling that is not coupled to the polyglot persistence choices we make at the individual service level. Then we will look at horizontal scalability and sharding so that we can understand the implications that this physical model has on our operational data modeling. Then, we will go through several examples of applying the single table design technique to the services in our hypothetical food delivery system.

Nodes, edges, and aggregates

For our operational...