Our recent poll results are in and have been tabulated. Following a discussion with Andi Mann (@AndiMann), we agreed the results were surprising: which cloud does not matter. At the same time, some of the responses confirmed a few thoughts that we have espoused for years.
We posed one question to the readers of our site, who are from all over the world and tend to be cloud and virtualization savvy: Which Clouds Do You Use?

What we found was that the type of cloud we use really does matter and that we are not hypervisor agnostic yet. We also learned that OpenStack has a small following and is not employed for general use for production workloads. Now, some may argue with these conclusions, but given who our readers are (cloud and virtualization savvy IT professionals, managers, and C-levels), we are willing to stand by the results. Here are the general rankings:

Amazon Web Services: 39%
Microsoft Azure: 35%
Private VMware vCloud: 27%
Not using IaaS cloud: 18%

These are the top rankings for the poll. What’s intriguing is the percentage of folks who are not using IaaS-based clouds. Also interesting is that Amazon and Azure place higher than private clouds of all types.

VMware vCloud Air: 14%
Rackspace: 12%
OpenStack (any form): 12% (8% private and 4% public)
IBM SoftLayer: 4%
Public VMware vCloud provider: 2%

Also of interest was the number of respondents who chose more than one cloud option (36% of respondents):

5 Clouds: 4%
4 Clouds: 6%
3 Clouds: 10%
2 Clouds: 16%
1 Cloud: 64%

While this reinforces the belief that the type of cloud does not matter for some, the response regarding VMware vCloud Air demonstrates that the hypervisor still matters for others. Thirty-six percent are using a cloud of clouds, which is what many clouderati would expect. I personally did not expect people to use more than two clouds. This puts cloud in a very different light. If we look at that 36% of respondents, it is equally interesting to see that 50% use just Amazon and Azure together.

Using Amazon and Azure: 50%

It looks like clouds are being chosen for specific jobs based on their requirements, not on the need to centralize management to one set of tools. This is very similar to the vi vs. Emacs wars early in the development of Unix, or the shell scripting scuffles early on in the Linux world, or even the browser wars. People predominantly use what they know, or if coming in with no preconceptions, pick the editor or scripting tool that fits their needs. The same applies to clouds. Not all tools are the same, and some are particularly well suited for specific tasks. When organizations choose clouds, they seem to be following the same patterns—they’re picking clouds based on what they’ve used before, or to meet specific task requirements. The impetus for cloud adoptions doesn’t seem to be driven by a need to standardize and simplify their IT. This is further confirmed by the fact that of the 36% that are using a cloud of clouds, 50% are using Amazon and Azure: two very different clouds with two very different management suites.
Additional useful poll information included the geographic distribution of our respondents. We had expected the majority of The Virtualization Practice readers to be located in the US. They were, but just barely. The distribution was pretty close to 50/50.

US: 51%
GB: 15%
NL: 6%
DE: 6%
Rest of world: 18%
AU: 4%

There is even more surprising geolocation data. If we look at the United States and Europe, we can come up with the percent not using IaaS, but using public and private clouds. The US and EU numbers are fairly close, and the difference could be attributed to the need for more compliance within the EU. I would have imagined there would have been more public clouds in use, given the number of VCE Vblocks and other converged environments claimed to have been sold for private cloud use. It could also be that they do not see these as clouds, but rather as just virtual environments.

US: None: 11%, Public: 71%, Private: 18% (45% using Amazon Web Services or Microsoft Azure)
EU: None: 13%, Public: 64%, Private: 23% (37% using Amazon Web Services or Microsoft Azure)

So, what real conclusions we can take from such a simple poll, once more taking into consideration that our readers are fairly cloud savvy?

  • Amazon and Azure are pretty close competitors, which explains why Amazon is pushing out new features to better compete with Azure, such as Amazon Workplace Email, a centralized Identity Access Management (IAM), and other features using Amazon’s Directory Services. We wrote previously that Active Directory is Azure’s secret sauce.
  • Clouds are chosen for features first, not for standardization, and a cloud of clouds is extremely popular. This may be caused by shadow IT and the fact that all those shadow IT tasks are being brought forward into the lIghT for management. With 50% of those using a cloud of clouds using both Amazon and Azure, the case can be made that there is a lack of standardization.
  • The hypervisor still may matter for some workloads, given the numbers for VMware vCloud Air. This is an avenue by which VMware private clouds can easily expand past their data center boundaries.
  • Geolocation does not really matter, either. Those who wish to use the cloud do so regardless of geolocation, but they choose clouds that are blessed by the authorities.
  • The percentage of private clouds in use is a surprising number. You would think it would be much higher, but organizations choose the bigger public clouds over almost everything else.

Final Thoughts

The amount of data we can get out of even a single-question poll is pretty surprising. We now have a good sense of the use of private and public clouds within Europe and the US, and we can draw some interesting conclusions about why clouds are chosen. With many organizations choosing to use a cloud of clouds, we know that clouds are being chosen for functionality over standardization. This may imply that we need more tools that can move data between clouds for data protection, cross-cloud security policy management, and cross-cloud management.
We can also draw a completely different conclusion: that IT is not involved, that most clouds are chosen as part of shadow IT, and that when they are brought into the lIghT, IT is left with an inherited management nightmare.
In either case, we do know that clouds of clouds are in use. We also know that clouds provide the necessary privacy to meet EU guidelines; otherwise, they would not be in use there. Further, we see that the big winners are Amazon and Azure.
Our goal is to re-run this poll in six months and track the changes. We may find that the trend is not what we thought. Were you surprised with these results?

2 replies on “Which Cloud Does Not Matter”

Comments are closed.