Considering the prevalence of cloud computing among businesses today, many held the belief that dedicated or baremetal servers will soon lose favour as an IT solution. Not only has this proven to be incorrect, but there has in fact been a growth in dedicated server use worldwide. In this article, we discuss the reasons why.
As you can see from the graph below (courtesy of Statista, an online statistics and research company), the global demand for dedicated server hosting services remains strong netting more than $5.72 billion in 2018. By 2020, the worldwide market value of dedicated server hosting is projected to reach as high as $6.53 billion.

But why? Isn’t the cloud the future? 

Not necessarily. The most likely factor for this pattern is that many organisations are extremely tech-oriented, which intense computing requirements (think of Software-as-a-Service vendors, gaming industries and the like). Almost all of these organisation want complete control over their IT infrastructure – something that the cloud does not always provide.
Hyper-scale cloud hosting is designed for businesses requiring rapid scalability: if your organisation frequently has to deal with unexpected peaks in server activity and requires automatically scaling environments, then the cloud is undoubtedly the ideal solution.
However, what if your organisation has applications that require a lot of horsepower, low latency or every last bit of CPU performance? In this case, you will probably see your cloud costs soar, as most hyper-scale providers charge a premium for dedicated CPU cores or higher frequency compute environments. Even then, the multi-tenancy of the cloud (where multiple different businesses are hosted from the same physical server) can degrade performance. Such cases highlight the reasons why dedicated servers will remain relevant for the foreseeable future.
Under certain use cases, dedicated servers are more efficient and cost-effective compared to the cloud. It is worth sticking with or moving to dedicated servers if your organisation falls under the following categories:

• Your cloud spending is large and reaches more than 500 AUD per month or more.
• Your organisation run applications that require optimal network performance with the least amount of latency possible. Ideally, you would want a hosting service that offers dedicated network capacity.
• Your applications and software demand high processor and memory performance. Many modern applications require rapid and predictable responses from CPU and memory to deliver the expected performance. This is often the case for organisations that maintain large databases running in memory.
• You have an existing base compute requirement that does not peak and trough, meaning you do not see much benefit from the ‘pay-as-you-go’ model of hyperscale cloud packages.