Capacity Planning and Architecture - Expert Oracle RAC Performance Diagnostics and Tuning - page 29

Figure 2-2 shows a high-level system stack and what is involved in a user request and system response overhead.
Figure 2-2. Oracle single-instance and cluster stack (PDB = pluggable database)
If you apply this to the database tier and expand the components of a server in a database stack, the stack would
have a resemblance to Figure 2-2 for a single instance and an Oracle clustered configuration. This means there are
several more measurements from various processes in a database stack that would need to be applied to our previous
capacity planning equations. In other words, when a user makes a request for set of data, the request is not the only
session or operation that will be on the server. There are other processes that manage the server and data that would
be involved in the operation. Figure 2-2 compares the stack in a single-instance and clustered configuration.
To understand the stack and the overheads involved, it's important to understand the architecture of the stack and how
data is processed and what kind of bottlenecks can affect the overall performance and capacity of these servers. In the next
few sections and chapters ahead, we dissect many of these components and understand the overheads and optimization
techniques. When deciding on the capacity of the servers, it would be good to take all this into consideration.