Cloud Server CPU Comparison

Comparison: Cloud Server CPU

The core of each Cloud Server, or Virtual Machine as you might call it, is the CPU.
Processors manufacturers such as Intel and AMD have their "own" method of measuring the performances of their processors, however you can base your comparisons of a physical server's CPU model according to the data provided by the manufacturer, so that you can compare two offers of servers in colocation/housing.

On the other hand, that isn't possible in the Cloud industry as no standard definitions of what a vCPU is exist!

Many Providers, like Amazon and Microsoft, offer shared CPUs.

Other Providers provide a dedicated computing power based on allocation and reservation; in the case of a dedicated CPU, several times it may remain inactive or slightly used for most of the times.

But shared CPUs are victim of "bulky neighbours", other Virtual Machines on the very same physical Host that monopolise the computing resources of the CPU.

If we consider the utilisation ratio of Cores adopted by most Providers, 1:8, you can understand how, potentially, for each Core assigned to your VM, there are at least 8 other Cloud Servers that compete for the utilisation of your CPU. Similarly, if you have a 2 Core cloud server, you might find out 16 other virtual machines (with 1 Core each) to challenge for the computing resources of your 2 Cores.

If this situation worries you, you should know that a dedicated CPU can guarantee secure and constant performances through time.

Cloud Services that offer dedicated physical Cores make easier the job of monitoring and predict performances, and can carry out the required work loads.

If then, in the end, Cloud Providers virtualize the CPU's Cores in different method, it gets incredibly difficult to find the right amount of CPUs for a specific application.

The variety and the inconsistency of performances of the shared CPUs call for more difficulties in predicting performances.

2017 Stellar is a CoreTech Srl service. All rights reserved