StackExchange Architecture Updates - Running Smoothly, Amazon 4x More Expensive
Monday, October 24, 2011 at 9:01AM
Todd Hoff in Example

We've had a few articles on the StackOverlflow Architecture and Stack Overflow Architecture Update - Now at 95 Million Page Views a Month. Time for another update. This time from a podcast. Every week or so Jeff, Joel and guests sit around and converse. The result is a podcast. In a recent podcast they talked about some of their recent architecture issues, problems, and updates. And since I wrote this article before my vacation, they've also published a new architecture update article: The Stack Exchange Architecture – 2011 Edition, Episode 1.

My overall impression is they are in a comfortable place, adding new sites, adding new features, making a house a home.

Notable for their scale-up architecture, you might expect with their growth that they would slam into a wall. Not so. They've been able to scale-up the power of individual servers by adding more CPU and RAM. SSD has been added in some cases. Even their flagship StackOverflow product runs on a single server. New machines have been bought, but very few.

So, the StackOverflow experiment shows the scale-up strategy for even largish sites is a good practice.  True, their product naturally separates by topic, much like the early Facebook, but Moore's law and quality engineering are your friends. They estimate Amazon would cost them 4 times much.

Here's what StackExchange has been up to:

StackExchange is refreshingly open. They consciously try and make their issues public by posting in their blog or asking questions on one of their sites. Maybe your organization might want to follow their example? It's an awesome way to build awareness and trust.

Article originally appeared on (
See website for complete article licensing information.