Checkout the slides

Transaction isolation levels are officially defined in terms of which data anomalies they prevent. However, terms like “phantom read” are difficult to understand and are not helpful to most application developers trying to choose the right level. Fortunately, there are other ways of looking at isolation levels that can remove the confusion from this choice.

In this talk I will describe how the choice of isolation level interacts with the design of the application. Practical differences between the levels will guide you to the isolation level that is right for your project. I will also discuss similarities and differences between different database implementations including PostgreSQL and CockroachDB.