While running some hardware benchmarks on my new Core i7 Apple MacBook Pro, I got the idea to run the same benchmarks across several Cloud hosting platforms to see how their hardware stacks up. The tool I used to perform the benchmarks is Geekbench from Primate Labs. Now, this benchmark was designed to accurately test processor and memory performance so these benchmarks do not take into account disk i/o, network i/o, etc. In addition to the performance results, we learn a bit about the underlying hardware on the host machines that run these platforms. The different providers benchmarked here are: Amazon EC2, GoGrid, Linode, Rackspace Cloud Servers and Storm on Demand (by LiquidWeb).
So about a week ago I got very impulsive and dediced to upgrade from my 3.06 C2D MacBook Pro to the new Core i7 based model that was recently released. I wasn’t simply after the latest and greatest processor however, there were some other significant changes that I was eager to experience first hand, so this is the model I purchased:
- 15 Inch Matte High Resolution Display (big selling point)
- Core i7 2.66 GHz
- 4GB RAM
- NVIDIA GeForce 330M 512MB
I really wanted to purchase a laptop with some newer architecture than the Core 2 Duo chips that have been out for a while as I’m looking to stay with the same machine for at least 2-3 years, which if you know me is near impossible! However, I finally brought myself to purchase AppleCare now that I’m comfortable with a machine that will last me for a while.
Strong web site performance, reliability, and uptime – these are all traits you would expect from an enterprise-class Cloud hosting platform, right? That is exactly how I feel now – and exactly what I expected almost three years ago when I started looking at the very early cloud hosting platforms of the time.
My name is Chad and today I am a Mosso employee. Long before I joined the Mosso team, however, I enjoyed (sometimes) hosting many of my own web sites and those of friends, colleagues, family, etc. Back in late 2006 I was looking for a better platform for the sites I hosted. I was tired of trying to manage my own solution, and I was frustrated with the downtime and poor support of all the VPS providers I had used over the years. Cloud hosting was a very new concept and, in fact, I don’t think it was even referred to as “cloud” just yet, but there were two main providers in this space at that time: Mosso & Media Temple. Like many, I was concerned about the initial concepts but very intrigued with the technology so I splurged and purchased hosting accounts both with Mosso and with the Media Temple Grid platform. I decided to do some testing on two basic fronts: response time & uptime. As you may have figured by now, one platform excelled in comparison, and I was so intrigued that I ended up coming to work for Mosso!
So now, in early 2009, I thought, “Why not duplicate those same tests and see what the results would be?” And that is exactly what I did. I set up identical static web sites serving a 100KB image and identical dynamic sites, a vanilla WordPress 2.7.1 install on both platforms (Cloud Sites from Mosso and Grid-Service from Media Temple). Next I decided to leverage Pingdom for all the testing, as they have a wide variety of testing locations around the world. These tests for uptime and response time were performed via a full HTTP check (not ping) and occurred in 1 minute and 5 minute intervals on different copies of the static and dynamic sites. I set the test to run during the middle of a typical business week and collected roughly 72 hours of data.
The results were very impressive. Uptime percentages for Mosso Cloud Sites were a solid 100% compared to about 99.0% on average with Media Temple GS, and downtime stuck solid to zero seconds, while Media Temple was down 1-5 times across the different test sites resulting in almost 30 minutes of downtime on average per site. Average response times for Mosso Cloud Sites were at least half of that of MT GS, with Cloud Sites sometimes operating almost as much as five times faster!
This test is easy to duplicate if you would like to try it firsthand. If you have any questions, please do not hesitate to contact me. Please see the results below and you can see the direct Pingdom reports here: http://www.pingdom.com/reports/ndv2m2bh6fnw/