Queries related to Neo4j best practices

Neo4j Graph Platform
1

Hi,
I have following questions in Neo4j 5.7.0 architecture.

  1. In the composite database example, why do we call regional customer data as different databases db1 and db2? Why did we partition the customer data to two databases, it can be sharded into different servers specific to the original region know.
    Ideally shards means distributing large database across multiple machines right.
  2. Why do we need alias in Composite databases? It seems similar to the relational database, can't we directly use the database name or does alias have some logic behind the query
  3. Is there any best practices for deciding the number of nodes in the core servers based on the number of shards or based on the data?

Thanks in advance.