Tidy Cloud AWS Issue 3 - Why Hashicorp's state of the cloud report raises more questions than answers

Recently, Hashicorp released their first State of the Cloud report.

The respondents to the survey were a bit over 3200 from the Hashicorp email list (out of 300000). These included people from all over the world, and from both small and large companies.

It is no surprise that these results point in the same direction as the views and goals of Hashicorp. After all, you are at the very least interested in their take on infrastructure management and tooling if you are on their email list.

Nevertheless, the report is an interesting read, but not from a perspective you would expect.


An important takeaway in the report is that multi-cloud adoption is a reality for many companies.

What does multi-cloud mean, though?

My interpretation is that it is about using multiple cloud providers within the same company or enterprise. That is very different from trying to run the same stuff transparently on multiple cloud providers. Instead, different workloads may run in multiple clouds - pick the best cloud provider for the specific job at hand.

In that sense, a lot of companies are multi-cloud today. For example, a company may run all of its core business solutions in AWS but have their office solutions with G Suite (Google Cloud Platform), or perhaps Office 365 (Azure).

Other examples would be to run the solution infrastructure in Amazon Web Services, but use Azure DevOps for CI/CD pipelines.

Many companies also use various SaaS offerings, which may run in one or multiple clouds. In some cases, you know what their cloud provider is, sometimes not.

If multi-cloud means use best of breed rather than use the lowest common denominator, then I agree that multi-cloud is a valid and possibly useful reality.

However, Hashicorp does not define what multi-cloud means here. Without a clarification, some results may be of questionable value.

Cost concerns, security concerns, and (lack of) skills are inhibitors for cloud work

In particular, in the context of multi-cloud, it is no surprise that cost, security, and lack of skills is a concern.

Security will be more complex where multiple cloud platforms are involved, in particular for workloads that cross cloud boundaries, and persons working with multiple cloud environments. It is simply more complex than for a single cloud platform. Even a single cloud platform in itself can be complex. How much of a concern is this depending on the approach to multi-cloud, or compared to single-cloud-oriented companies?

Cost can be both a driver and a concern. Is cost a concern due to the multi-cloud approach? Or have cost concerns been the driver for a multi-cloud strategy? It can be both, and the report does not answer that.

Lack of skills is also no surprise. To understand a single cloud platform is challenging and multiple platforms? Very tough! An understanding of multiple cloud providers in-depth? Very few have those skills. Is this concern reflecting the specific subset where a person has in-depth multi-cloud skills or is this a concern about in-depth cloud skills in general?

Again, I do hope these concerns and inhibitors are there for the right reasons and that it is about managing best of breed. Or perhaps, meet the customers where they are if you run a SaaS business with cloud services.

Automation and infrastructure-as-code is king

A vast majority of the people responding to the survey consider automation and infrastructure-as-code either extremely important or important for the success of the cloud initiatives.

My initial reaction to this was more or less like this:


I do believe this is very important. Automation is one piece of the puzzle to get reliable, consistent, and repeatable solutions.

However, it would not be a surprise to think that 1% of the people on Hashicorp’s mailing list thought that. I suspect that a different demographic would give another result. Many people would think more about outcomes - and so they should.

For most people, automation, and infrastructure-as-code should be there behind the scenes and just work. For those of us that are more deeply into that area, it is our job to make it so.

Final notes

I think the report was an interesting read, but for me, it raises at least as many questions as it answers.

For Hashicorp I think it probably provided some possibly useful feedback, but more than that it provided them with a nice marketing tool.

If this interests you, check out my website at Tidy Cloud AWS. It contains both AWS-specific material, as well as more cloud-agnostic material.