I've spent the last few days I've spent some of my free time looking at an Infrastructure as a Service platform named Heroku. I'm probably just one of many people who've written similar posts to this, but what's the point of having a blog if you're worried to say what's on your mind.
I've setup a few simple projects so far. It's insanely easy to get started. Which is great. The part that concerns me about it is the financial viability of hosting a large site there. Past about the $50/month range (two dynamos and a 20GB postgres instance) it seems that the time cost of migrating from Heroku to a service such as Linode is likely well worth it.
Looking at some of the specifics of the implementation it seems that the Heroku is really just a very easy to use wrapper around Amazon Web Services. Which is really neat from a technical achievement perspective, and also explains the pretty aggressive cost scaling:
I think I'm going to follow this up with more posts. It's been a lot of fun to get some simple projects up and going and I'm trying to spend some more of my personal time writing code again. I'm doing more and more IT/SysOps at Curse over the last few years and I think that long term I'll be doing that full time. While I am a fan of DevOps it's something that still only limits it to so much code, and I don't want to get too dull.