One of the main reasons why organizations flock to public clouds such as Amazon Web Services is that their pay-as-you-go infrastructure services are seen as more economical than buying expensive on-premises gear. But according to a newly published report from 451 Research, this is not necessarily the case for every company.

The paper is the latest in a series of quarterly studies from the firm that assess the cost-effectiveness of the various cloud solutions on the market. It claims that while infrastructure-as-a-service platforms are the least wasteful of the options out there because they allow companies to provision capacity according to their demand, there are certain conditions where on-premises deployments can be more economic to operate. In particular, 451 Research points to large data centers with high labor efficiency.

The firm’s analysts found that private clouds generally provide a lower total cost of ownership than the public cloud if there are more than 400 virtual machines per engineer. Owen Rogers, the research director for 451 Research’s Digital Economics Unit, explained that “salaries and labor efficiency have a disproportionately large impact on pricing” in enterprise information technology departments. The fewer techies a company needs to hire, the less it will have to spend on operating its infrastructure.

But today’s report highlights that there are also numerous other factors involved in the cost-effectiveness of a cloud deployment. Despite the fact that there many more administrators trained in operating VMware Inc. and Microsoft Corp. environments than OpenStack clusters, 451 Research found the latter to have a better TCO when there are more than 400 virtual machine per engineer. And the platform can become even more cost-effective when companies use commercial distributions such as Mirantis Inc.’s that layer various automation tools on top of its core functionality.

At the same time, however, Rogers stresses that companies should take more than just cost-effectiveness into account when choosing a cloud solution. If possible, he recommends that they “consider a hybrid or multi-cloud strategy so they can determine the best execution venue for each workload based on cost, management, technology and location requirements.”

Image via Pixabay