If you have different databases, each one has been created to manage a certain type of transactional data very efficiently. One has been created to manage application and accessory data efficiently. But they both have to communicate with one another. You have duplication data, and it has to be shared back and forth between those two databases. That means you've basically doubled the amount of storage you need, and you've reduced the transactional time because it has to talk to each other before it can talk to you. That's what the technology eliminates: the duplicated databases. You can have multiple databases but only one instance that is used globally, eliminating a lot of replication.
This version of SAP came out in 2015, and everybody's moving to it because they're not really supporting the old version.
You can pick one of the hosted cloud services as opposed to owning it and doing it yourself. Smaller companies cannot buy a bunch of big servers and storage devices and hire people to run the data center. They want somebody else to do it. SAP is also a multi-tenant, so SAP will work with AWS, Azure, and IBM.
If you're one of the tenants in a big server and don't want people to see your data, you can have your own database separate from customer B and customer C. There's a price for that. That's basically private. There's a certain kind of cloud hosting where you're the only tenant. For government companies that are high-tech and don't want their data stolen, they tend to do a private solution, which means they're the only one using it.
You can also use this as a hybrid solution, which allows you to be a tenant in the solution. Your data could even be shared, but company A's data can't be seen by company B or C and vice-versa. That's a public solution where you're using the same database but different people can't see your data in that database.