Tobias Petry

Scaling Databases

Stop over-engineering your database. Learn the simple scaling patterns that can handle massive traffic before you ever need to shard.

Scaling Databases
#1about 2 minutes

The database is the real bottleneck in scaling applications

Your programming language or framework is rarely the performance problem; the database is almost always the component that limits growth.

#2about 1 minute

Optimize queries and add caching before scaling out

Before distributing your database across multiple servers, you must first fix slow queries, add proper indexes, and implement caching to avoid amplifying existing problems.

#3about 1 minute

Scaling up is the simplest and most effective first step

Instead of immediately adding complexity with multiple servers, simply upgrading your existing database server with more memory and CPU cores is a cost-effective solution.

#4about 4 minutes

Understanding the trade-offs of multi-master replication

While multi-master replication allows writes to any server and guarantees read-after-write consistency, it often suffers from performance degradation due to write conflicts on the same data.

#5about 5 minutes

Using read replication to scale read-heavy workloads

Read replication uses a single primary for writes and multiple secondaries for reads, but requires careful application design to handle asynchronous replication lag and potential stale data.

#6about 5 minutes

The power and complexity of database sharding

Sharding provides near-infinite scalability by distributing data across multiple independent databases, but introduces significant complexity like cross-shard joins and managing multiple database systems.

#7about 2 minutes

Keep your database scaling strategy as simple as possible

Avoid premature optimization by choosing the simplest scaling solution that meets your current needs, as overly complex systems are difficult to manage and can lead to costly rewrites.

Related jobs
Jobs that call for the skills explored in this talk.

Featured Partners

Related Videos

Scaling: from 0 to 20 million users28:12

Scaling: from 0 to 20 million users

Josip Stuhli

Database Magic behind 40 Million operations/s24:22

Database Magic behind 40 Million operations/s

Jürgen Pilz

Fault Tolerance and Consistency at Scale: Harnessing the Power of Distributed SQL Databases32:46

Fault Tolerance and Consistency at Scale: Harnessing the Power of Distributed SQL Databases

Wei Hu

Fault Tolerance and Consistency at Scale: Harnessing the Power of Distributed SQL Databases28:41

Fault Tolerance and Consistency at Scale: Harnessing the Power of Distributed SQL Databases

Wei Hu

Single Server, Global Reach: Running a Worldwide Marketplace on Bare Metal in a Cloud-Dominated World50:06

Single Server, Global Reach: Running a Worldwide Marketplace on Bare Metal in a Cloud-Dominated World

Jens Happe

How building an industry DBMS differs from building a research one30:34

How building an industry DBMS differs from building a research one

Markus Dreseler

Maximising Cassandra's Potential: Tips on Schema, Queries, Parallel Access, and Reactive Programming31:08

Maximising Cassandra's Potential: Tips on Schema, Queries, Parallel Access, and Reactive Programming

Hartmut Armbruster

Branch your database like your code: How schema changes and pull requests go hand in hand48:05

Branch your database like your code: How schema changes and pull requests go hand in hand

Johannes Nicolai & Lilli Seyther-Besecke

From learning to earning

Jobs that call for the skills explored in this talk.