Speaker: Jim Walker

Distributed systems and distributed thinking are becoming more popular as we look to squeeze efficiencies out of our once wasteful monolithic applications. The cloud promises elastic scale and can give us an always on and available infrastructure, but our approach to building apps may need to be rethought to take advantage of this promise.

In this talk, we will explore the architecture of CockroachDB, a naturally distributed database designed to deliver transactions at global scale. We will walk through some of the design considerations we considered when architecting our database and provide some key best practices.

How do you design for scale and resilience IN your application instead of around your application?

Some of the topics we will cover include:

• Designing for scale and resilience
• Orchestrating distributed services
• Distributed principles and application design.

There is no reason we should be coding for data quality in our apps.

Is your data correct? Do you “get” isolation levels in a database and why they are important? Do you think in referential integrity? Foreign keys? Every app has different requirements for data correctness and the power of a SQL can help you manage these things.

The relational database is powerful but misunderstood by many. And the concept of ACID transactions is used loosley without much depth of explanation of how it is used. In this session we will dive into relational SQL and talk through:

• What is data isolation and what can go wrong with your data
• What are isolation levels for a database
• Dirty Reads, Phantom reads, Non-repeatable reads, and the problems with inconsistent transactions
• Referential integrity, foreign keys and the power of SQL.