As an architect, I would like to rework the flow of how data is accessed from individual services.
We need to review the reusability and maintainability of our code base in order to implement access to an external ORM database. We could use either use a pratical SQL server or a NoSQL implementation. Either could work, but that would depend on which would scale better.
Database tables are painful to migrate, if we ever need to do so, even with modern tools that are supposed to ease the complexity of having minimum data loss high.
If I recall, MongoDB, a NoSQL server, boast being performant and highly scalable. If this is true, we would only be limited to hardware performance. On the other hand, most popular SQL servers are just as performant and scalable; this might just be a wash between the two.
Certain NoSQL servers can, or are able to, consume JSON objects as-is. In this case, our need to implement DAOs would simplify considerably, removing any need for massive dependencies and/or need for shared libraries to be installed our machines to interface. This includes requiring little complexity of translating our from database to application instances.
As an architect, I would like to rework the flow of how data is accessed from individual services. We need to review the reusability and maintainability of our code base in order to implement access to an external ORM database. We could use either use a pratical SQL server or a NoSQL implementation. Either could work, but that would depend on which would scale better.