Build vs Buy

I’m catching up on all of my favorite blogs right now while I’m still on vacation.

I’m really enjoying great posts about cloud computing from Fred and Alex.  

When it comes to basic web infrastructure, startups have a variety of build vs buy decisions. Bandwidth is always a buy decision so that’s not really a choice. 

But there are various choices when it comes to storage and computing. Build vs buy. Hardware is cheap but that’s only part of the true cost.

Amazon Web Services has been a great choice for many startups. Most of the ones I’ve seen up close use S3 and some are using EC2 to get started.

I agree that going with a big company like Amazon is a smart choice for this stuff especially in the beginning as user demand is unclear and the costs are low at first. And it’s easy to get up & running. And we are seeing alternatives coming online like Rackspace new computing service from Mosso. I’m guessing that Microsoft and Akamia will have similar services sometime in 2008. Maybe IBM will follow as well?

But at some point a startup goes to the next phase of it’s growth and at then they have to make another build vs buy choice. Cloud computing scales but at what price or cost?

The AWS Success Stories page lists early stage companies. I’d love to see a blog post or white paper that describes the costs of AWS or Mosso as a web service goes from 500k active users to 1M active users to 5M active users (and up) along with various assumptions on storage and computing requirements.