Firebird transactions in details

This course is devoted to Firebird transactions: how they are related to multi-versioning architecture, how the transactions isolations works, isolation levels, transactions markers, how Commit and Rollback works, garbage collection principles, sweep, and diagnostics of problems with transactions management in Firebird applications.
Please see the detailed schedule of the course below:

How record versions are being created in FirebirdSQL
States of transaction in Firebird (active, committed, rolled back, limbo) and what they mean
Visibility of record versions in transactions
Key differences between Read Committed and Snapshot
Triggers for transactions start and stop
Context variables at transaction level
What happens at transactions’ start
Global transaction markers (OAT, OST, OIT) and their meaning
What happens at transaction commit
Standard rollback with cache flush
Forced rollback without cache flush
Rollback transactions with undo-log
Rollback without undo-log
What is retaining transaction
Temporary BLOBs
Undo-log
Auto-commit (does anybody know about it?)
Managing memory for transactions
Local TIP and related issues
Transactions interaction and best practices for application development. Conflicts between transactions.
Templates for transaction usage in application development
Scenarios breakdown: Readers and Readers, Readers and Writers, Writers and Writers
Best practices
Garbage collection mechanism and sweep in details
Differences between sweep and garbage collection
Autosweep: pro and contras
Examples of application design mistakes which lead to excessive garbage
Diagnosing transaction management problems: most prevalent problems, how to catch and fix them.
OIT, OST, OAT, NEXT — their combinations and what they mean
When the garbage collection does not work?
How to determine quantity of garbage versions
How to find long transactions with MON$ tables
Transactions monitoring with MON$ and TraceAPI.
Legacy applications: workarounds for error in transactions management