Never been one to let the carrier drop

So, I’ve deployed many a application to EC2 (amazon’s cloud service)

I think that the good things about it are well covered on Amazon’s pages talking about it. There are, however, some bad sides they don’t mention:

1) Amazon is not subject to Moore’s law. In the time one of my customers has been on there, CPU power has quadrupled for the price and so has memory, and disks have improved by two orders of magnitude thanks to cheap and reliable SSDs. However, Amazon continues to charge the same price they did three years ago, and does not offer a SSD at all. They offer the option to purchase a specified number of IOPS, but the price is awful – you could buy a SSD every month for what they’re charging for a mere 1000 IOPS.

2) Amazon is not as reliable as one might like. Several times we’ve had instances get wedged in ways that required filling out a ticket and waiting several hours to repair. And today, every single EBS volume went down at 11 AM, and half of them are not back and it’s already 3:30

3) There’s no one to call. With managed hosting or colocation there’s usually someone you can get on the phone when everything is breaking. At amazon, all you can do is fill out a feedback form, and you don’t even get a ticket # or a response.

4) If a disaster does happen, everything gets very sluggish because everyone is trying to get into the web interface to fix things at once, and everyone is rebuilding their servers from AMIs at once.

5) Disk space at amazon is *expensive*. You basically buy the hard disk every few months.

This entry was posted on Monday, October 22nd, 2012 at 3:28 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.