Database Reference
In-Depth Information
Ideally one would combine the lower memory and lower process footprint of the multi-tenancy RAC database
and the application isolation of the single database per host. Well you may have guessed it—that's what the Pluggable
Database will offer: horizontal virtualization, but from inside the database. PDBs offer the best of all worlds, and allow
you to achieve a much higher consolidation density if your hardware supports it.
Pluggable Databases to the rescue
With Oracle 12.1 a fundamental change has taken place in the architecture. Luckily, as with most technology introduced
in the database, you still have the choice not to use it straight away. For those who do, and I hope I can encourage you to
do so with this topic, you will have the option of using a new database type: the Container Database or CDB.
The main benefit of using the CDB is to allow for horizontal virtualization within the database. A Container
Database, as the name implies, contains at least one, and up to 253 user-Pluggable Databases. The Pluggable
Database is the consolidation target. From a user's point of view a Pluggable Database (PDB) appears just like any
normal database before Oracle 12.1. As such, a PDB is a collection of schemas, tablespaces, and other metadata
objects. The Container Database itself is merely a container and does not store user data.
The Container Database serves two purposes. First, it is the container for all these Pluggable Databases you are
about to create. Second, it contains what is referred to as the CDB's root. Named CDB$ROOT in the documentation it
is the home for information common to the CDB and the PDBs. The CDB has a further component named the seed
PDB or PDB$SEED. The seed database is a minimal PDB that you can use like the starter database when creating a
database using the Database Configuration Assistant dbca. It cannot be opened or modified.
The Container Database instance type exists alongside the database type as we know it from pre-12.1 days.
In the remaining chapters I will refer to this former database type as a “non-CDB” to make a clear distinction between
the two. Interestingly the Container Database is not limited to single instance deployments, it can also be created as a
Real Application Cluster database. The deployment does not have any impact on the Pluggable Databases for which it
provides a home. Consider Figure 7-2 , outlining the new database “container database” instance.
Figure 7-2. A comparison of CDB vs. non-CDB
 
Search WWH ::




Custom Search