IMPORTANT: This article applies only to our legacy VPS plans and not any of the new cloud options.
As part of the mvScale virtualization system, our engineers use a range of complex algorithms to ensure that our customers receive the maximum processing power possible within the constraints of their VPS service package. Each package’s CPU allocation is based on the natural ratios created by the package’s memory allocation (both guaranteed and burst), but because our VPS accounts are deployed across an array of server and chip set types, there is no hard number such as mHz that can be used to define “how much” CPU is used. Instead, we use a unit of measurement called a “virtual processing unit,” or “VPU” to describe the relative amount of processing power allocated to each package, as follows:
VPS Small = 15 VPUs
VPS Medium = 40 VPUs
VPS Large = 60 VPUs
This cannot be specified in terms of Ghz or CPU percentages as the VPS hosting packages are spread across a range of server hardware and these metrics simply don’t work. Also, simply specifying a clock speed would not define processing power. Clock speed is only one of many specifications that go into what makes a CPU “fast.” Caching, die sizes, bus optimization, hyperthreading, turbo, and the like are all critical factors as well. Our mvScale algorithms tweak literally dozens of different parameters to ensure that each customer has the most processing power possible on any given package.