High Scalability

What is in the archive

High Scalability was started by Todd Hoff and became one of the best-known sources of real architecture case studies on the web. The archive includes:

• Company architecture posts: how named companies built their systems, often based on talks and interviews with their engineers
• "Stuff The Internet Says On Scalability" roundups of interesting links, published for many years
• Explainers: consistent hashing, gossip protocols, Kafka basics
• Outage analyses: for example a post arguing that the Swedbank outage showed the limits of change controls
• Scaling histories: a brief history of scaling Uber and a look behind Amazon S3's scale (both 2024)

The most recent post on the site is from May 2024, and posting was already infrequent from 2022. The archive remains online and free.

Why it is still worth reading

Interview designs are clean; real systems are not, and the gap is where judgement lives. Reading how a company handled a migration under load, or why it abandoned an approach that looks correct on a whiteboard, teaches something no worked example can. Older posts are often especially clear about reasoning, because they describe scaling problems that were new at the time and are now routine.

How to use it

• Search for a company or problem you are studying rather than browsing chronologically.
• Check the date of every post. A 2012 architecture tells you how a company scaled then, not how it runs today. Look for the company's own engineering blog for the current picture.
• Read for decisions and failures, not technology lists. Note what broke at each stage of growth and what they changed.
• Use the explainer posts as quick refreshers on classic distributed systems ideas.

Where it falls short

The site is effectively dormant, so there is nothing new to follow, and some links in older posts are dead. Some posts summarise talks rather than being written by the engineers involved, so details can be second-hand. For current architecture writing, company engineering blogs are now the better source.

Pair it with

• Awesome Scalability, which indexes many primary-source engineering posts by topic
• Company blogs in this library, such as Cloudflare, Discord and Dropbox, for current first-hand accounts
• Designing Data-Intensive Applications, for the concepts behind the case studies

Common questions

Is it free? Yes, the whole archive is free to read.

Is it still updated? Rarely. The latest post is from May 2024.

Are old posts still useful? Yes, for the reasoning and failure stories. Just do not assume a company still uses the architecture described.

Best for: engineers who want real-world scaling stories and are happy to read an archive rather than a live feed.