The next opening keynote – Everything fails, All the Time

Our third keynote this morning was by Dr Werner Vogels – CTO Amazon.com

His second question to the audience “How many of you don’t shop at Amazon?” When one or two people raised their hands he commented, “Can I talk to you later.”

He was here to talk about Amazon as the technology company, forget about shopping. Think about the technology that drives that. There were some large and impressive numbers about Amazon. The one stood out was 3.9M shipments for peak day (didn’t mention which day)

Some points from the session.

  • Some Big News — Persistence Storage for Amazon EC2
  • Amazon started as technology consumer, we are now a technology provider.
  • We were enterprise scale, now we are at web scale.
  • Taking the approach, get big fast, the result was every architectural principle was broken.
  • Moving into Software as a Service
  • Develop -> Test -> Operate
  • There is a box between Test and Operate (undifferentiated heavy lifting)
  • We seem to put data centers in the same places as trailer parks.
  • At Amazon we expect our data centers to fail.
  • We expect our software to tolerate these fails.
  • We want to be able to loss a data center and not have this affect our customers.
  • Amazon, brings data centers down multiple times a year, just to show the software survives.
  • The provider model — Scalable, Cost Effective, Reliable, Simple
  • Addressing Uncertainty.

He used the 365 main power failure example. Large Web 2.0 Web sites were down including: Live Journal, Craigslist, Technorati, TypePad, Yelp and USA Today.

Is Amazon a viable option in these times? Worthy of consideration.

Updated You can view more links including video courtesy of Sheeri Kritzer at MySQL Forge Conference Notes

Tagged with: Uncategorized

Related Posts

My take on PyCon Australia 2026

PyCon AU 2026 held in Brisbane this week was a five day event for the Python community. With specialized tracks and dedicated tutorial and engineering days there was a wide variety of content available.

Read more

Fixing Old Software Bugs Without Thinking

I previously came across an issue in my benchmarking work where I relied on the Latency Histogram section of sysbench output. On Linux it worked fine — I generally ran my workload there, so it never mattered.

Read more

Curating DuckDB Datasets Leveraging AI

In Curated MySQL Data Sets for Realistic Testing I described datasets assembled the manual way — download, schema, load, validate, document — over hours or days per source. This post is the follow-up I promised: what changes when AI assists the same workflow, using DuckDB as the target engine and GeoNames as the first example.

Read more