virtualization.info

One week ago virtualization.info published the results of its Virtualization Industry Survey 2008 about the Hardware Virtualization adoption.

Easy to guess, the responses for Q6 – What hardware virtualization platform do you implement? generated concerns among our readers, as Microsoft Hyper-V appeared “more adopted” than VMware ESX.
Some of you demanded a clarification on the methodology used to collect and validate the published data, others simply judged the whole survey as 100% useless.
A few noted that, differently from the usual behavior, there was no comment on the results. It was not a case.

Now that enough buzz was generated it’s time to disclose the real information we collected.

The responses were collected online, through a web form (developed with Google Forms) that accepts anonymous answers (so there is no profiling of the surveyed person) and doesn’t enforce any security measure (so any reader can vote multiple times during a single day or over a longer period of time).
The only tracker that the survey platform registered was a unique timestamp for each response form, reporting date and time (HH:MM:SS).

Some quick tests with Google Forms, that anybody can perform (as the product is free), would confirm that a visitor can complete the survey multiple times per day without being tracked or blocked.

Despite that, virtualization.info tracked every possible information about the people that visited the form page using a 3d party statistic engine.
For each visit we recorded things like the source network, the country, the city, the browser, and other typical information.
Of course this is not enough to unambiguously say how many from a certain source network submitted the survey form. But, depending on how much time a visitor spent on the page (and we tracked this information), we can guess with a good level of confidence if the visitor filled the form or not.

Before going any further, it’s worth to highlight that in an anonymous survey like this one, each vendor is in the same condition and can equally cheat to influence the results.
In an ideal world the vendors (or better, their employees) would recognize that this is a very bad idea, as their actions completely mess up the result and waste the efforts of the ones that prepared the survey and the ones that participated it with honesty.

Said so, the following numbers should be self-explanatory.

Oct. 29, 2008 – The survey is unofficially announced inside an article about the IDC and Gartner market shares reports.

Here’s what happened on the survey form page that day and the other two (Oct. 29-31) of that week:

• 237 page views from Microsoft Corp with 1:49 minutes per visit on average
• 147 page views from Microsoft European Internet Data Centres with 2:37 minutes per visit on average
• 473 responses marked “Hyper-V” as answer to question Q6

versus

• 75 page views from VMware Inc. with 2:30 minutes per visit on average
• 178 responses marked “ESX” as answer to question Q6

Here’s what happened on the survey form page over the entire survey period (Oct. 28 – Dec. 31):

• Microsoft (US and EMEA) page views – 551 with 1:38 minutes per visit on average
• VMware (US and EMEA) page views – 136 with 3:23 minutes per visit on average