Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

We've been looking at moving some or all of our stuff to either Amazon EC2/EBS/S3 or Rackspace cloud hosting, and it has been interesting.

Amazon seems more flexible, since you buy block storage (EBS) independent of instances. If you have an application that needs a massive amount of data, but only a little RAM and CPU, you can do it.

Rackspace, on the other hand, ties storage to instances. If you only need the RAM and CPU of the smallest instance (256 MB RAM) but need more than the 10 GB of disk space that provides, you need to go for a bigger instance, and so you'll probably end up with a bigger base price than at Amazon.

On the other hand, the storage at Rackspace is actual RAID storage directly attached to the machine you instance is on, so it is going to totally kick Amazon's butt for performance. Also, at Amazon you pay for I/O (something like $0.10 per million operations).

Looking at our existing main database and its usage, at Amazon we'd be paying more just for the I/O than we now pay for colo and bandwidth for the servers we own (not just the database servers...our whole setup!).

The big lesson we've taken away from our investigation so far as that Amazon is different from Rackspace, and both are different from running your own servers. Each of these three has a different set of capabilities and constraints, and so a solution designed for one will probably not work well if you just try to map it isomorphically to one of the others. You don't migrate to the cloud--you re-architect and rewrite to the cloud.



If you're interested to see how sites perform on EC2 and Rackspace over time:

https://www.blamestella.com/vendor/ec2

https://www.blamestella.com/vendor/rackspace


You're monitoring from AWS US-East it looks like, you'll want to mention that to give people some context around the latency numbers.


That true but I think what's more interesting is the number of incidents (timeouts, exceptions, and significant slowdowns).


I use the Rackspace cloud for a few Windows servers. The experience has been mostly positive, but they disappear for a few minutes each week, which is kind of troubling.


Disappear as in they crash and reboot, or disappear as in they're unpingable?


Gah, sorry for the late response, I just saw this. They just go unpingable for a few minutes. They seem fine when they return from purgatory.

It's always been at random times overnight, so far it hasn't happened when I know there's been load on the servers.




Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: