This document summarizes a presentation on Multi Version Concurrency Control (MVCC) in PostgreSQL. It begins with definitions and history of MVCC, describing how it allows transactions to read and write without blocking each other. It then discusses two approaches to MVCC - storing old versions in the main database (PostgreSQL) vs a separate area (Oracle). The rest of the document does a deep dive on how MVCC is implemented in PostgreSQL specifically, showing how tuple headers track transaction IDs and pointers to maintain multiple versions of rows.