Cloud Computing Comparison – Part 2

By | September 12, 2009

This is a continuing on from Part One of my comparison between the “Cloud” service providers Amazon and Rackspace



For the second part of my testing I ended up going with both the home page and the Apache Tomcat homepage.

These stats are based on an entire week with the days being August 21st  – September 6th

There did not appear to be any correlation to time of day and response time, the averages for the times of day were +/- .08 to the average of each hour of the day for each site.

The test is an HTTP GET request to the default Tomcat page on port 80 which contains 4 requests and 18KB in size on both the Amazon and RackSpace servers.

As a baseline, I tried to find a similar site in size and requests.
The baselines are:
Google search engine (5 requests and 34KB)
Apache Tomcat homepage (4 requests 18KB)

Source >1 second >2 seconds
Amazon         3.2 %           .74 %
RackSpace     3.2 %           .69 %
Google          4.2 %           1.0 %
Apache        11.4 %           3.5 %

All information is the percentage of all request based off time for page to load
Example (66 of 2021 samples were greater than 1 second total page load time which means 3.2% of all samples were greater than 1 seconds load time)

Next is the average page load times

Source Average (in seconds)
Amazon         .46
RackSpace     .43
Google          .44
Apache         .89

So when comparing the performance of the servers to other web pages of similar size and characteristics, it looks like we are right in line with where we should be. So what does all of this mean? I think it means that at least between these 2 providers, network speed, latency, general server performance and availability are nearly identical. Now I know to get a true test, this would need to run for months and months in order to more accurately analyze the performance, for my needs and my requirements(and time constraints) this was enough.

It hasn’t been asked of me just yet but I envision the need to run a formal stress test of the servers and once that is complete, I will post up my results from that as well. So there you have it! Take this information for what it’s worth. I wasn’t able to find much data regarding this before I did my testing which is why I posted it. I hope you find it all useful.

One thought on “Cloud Computing Comparison – Part 2

  1. Pingback: Cloud Computing Comparison – Part 1 – LazyNetworkAdmin

Leave a Reply

Your email address will not be published. Required fields are marked *