Hi
Sorry about the database version which I have mentioned while filling out this form.
This is a generic question, kindly consider it as for the latest stable release of 19c database.

As per my understanding a "connection" is a physical path and a "session" is more of logical. A session is associated with memory. And from V$PROCESS.addr shows the process id of a connection. If there is a connection then there is one entry in the v$process. Here a connection is getting associated with a process, and this process will be associated memory(PGA). Multiple sessions can use the same connection.
So I interpret this as multiple sessions associated with single process( process created because of connection). Since the process id is same the same PGA (associated with the process) is used. But sessions got session boundaries and transactional independence( considering a simplistic approach).
So can I please conclude there are multiple UGAs inside a PGA(dedicated to single process, spun as a result of the connection path)?
How the session level segregation happens inside PGA (single process dedicated) when a single connection used by multiple sessions ?
Is this data visible somewhere, like finding 5 UGAs are there in a single PGA?

Thank you.