Multicloud is becoming the de facto standard. Indeed, a solid 84 percent of the respondents in the RightScale report use more than four cloud providers, including both the public and private clouds. (Note, RightScale is now part of Flexera.) However, not only are companies shifting to multicloud, but to more than one public cloud as well. That means using Google, Microsoft, and AWS—two or three providers, typically, and sometimes more.
This was also shown in the RightScale report, with public cloud being the top priority, indicated by 31 percent of the respondents. Companies plan to spend 24 percent more on the public cloud in 2019 than they did last year.
The battle cry of multicloud is choice and the ability to avoid lock-in. Although choice is certainly a benefit, such as being able to pick best-of-breed cloud services, avoiding lock-in is not. You’re still writing applications using cloud-native systems, and by default that causes a lock-in to a particular public cloud platform.
What does this choice cost? I have a short list of issues that most don’t think about when going multicloud.
First, the cost of complexity. As stated here before complexity has a rather significant financial impact on security, operations, and governance. The more public clouds being employed, the more complexity. The more complexity, the greater the costs across the board, but mostly in operations.
Costs include tools to reduce the complexity, such as cloud management platforms and cloud services brokers, as well as additional staff needed for security operations (secops) and cloud management and operations (cloudops). The expense is typically 30 percent more in operational costs for each additional cloud, even when employing a sound toolset.
Second, the cost of increased training and hiring. Most organizations expect a higher price tag here, even moving to a single cloud. However, moving towards a multicloud, I’m finding that enterprises are often blindsided by the expense.
Consider the fact that multicloud people don’t really exist. Instead you have people with AWS skills and certifications, or with Microsoft or Google. You’ll have to hire two or three times what you would need with a single cloud provider, as a rule of thumb.
Many offer counterarguments about efficiency, but that’s perhaps less than a linear cost increase as the number of public clouds are used. I’ve not found that to be the case yet, so I’m sticking to my story.