You used to see them. Now you don't. Virtual CPUs dominate the cloud.

  • AWS, Google Cloud, Microsoft Azure, DigitalOcean and VMware all offer vCPUs.

  • Because vCPUs offer more efficient use of resources, Himanshu Singh, director of cloud platform and AI product marketing at VMware, said they are typically considered a better value for companies.

  • There isn’t really any type of workload that can’t run on a vCPU vs. a physical CPU.

Virtual central processing units (vCPUs) aren’t a new phenomenon in networks but they are becoming more prevalent. In fact, some experts say that vCPUs are so common that it’s more difficult to find use cases where it makes sense to use a physical CPU for a workload instead of a vCPU.

Roy Chua, founder of analyst firm AvidThink, said that vCPUs are a pretty standard offering from cloud providers. “It’s just a ‘virtualized’ instance of a CPU created via some form of virtualized technology like a hypervisor,” Chua said. Amazon Web Services (AWS), Google Cloud, Microsoft Azure, DigitalOcean and VMware all offer vCPUs. Smaller players, like IONOS, are also getting in the game.

Himanshu Singh, director of cloud platform and AI product marketing at VMware, said vCPUs offer enterprises many benefits because they allow them to pool their resources – using a combination of vCPUs and a physical server— and move workloads around. Singh added that, for example, some of these workloads might be running fast while others are running slow to accommodate different applications.

“A physical CPU could not handle everything. By virtualizing, you are able to pool and share resources,” Singh said, noting that this type of scenario gives enterprises better performance and more efficient use of their resources.

Because vCPUs offer more efficient use of resources, he said, they are typically considered a better value for companies. “You can get more utilization out of your investment.”

Singh said that VMware typically finds that most companies under-use their physical CPUs by as much as 50% because they aren’t able to move workloads and use all the existing capacity on their physical CPUs.

“Being able to move workloads around is huge,” he said, noting that many companies will use vCPUs to move their workloads from one vCPU to another just to do maintenance on the network and not risk having any downtime.

He also said that there isn’t really any type of workload that can’t run on a vCPU vs. a physical CPU. At one time, some experts believed that high-performance computing applications should be reserved for the physical CPU but Singh said that theory has been dispelled.

IONOS, a German-based IT firm that that recently started offering a vCPU to its small and medium-sized business customers, said that it recommends that high-performing applications use dedicated core servers where customers are able to select their virtual machine’s CPU model with each dedicated physical core.

Better security, cost control

Not only do vCPUs allow enterprises to more efficiently use their compute resources, the technology also provides better security because if there is a problem with a certain workload in the network it can be migrated to an area and isolated from the rest of the network, which prevents the problem from spreading.

IONOS claims its vCPU will help enterprises save up to 25% compared to similar offerings from the hyperscalers.

In a written response to Silverlining’s question about the cost savings, IONOS said that while public cloud prices vary across different data center locations, the company believes it offers cheaper prices per hour compared to the large public cloud providers like Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure. In addition, IONOS says it can provide less expensive traffic costs and administrative support than the bigger cloud firms.

Singh said that smaller companies that experience “sticker shock” when they start to look at moving workloads to the public cloud may find it more affordable to use their own data centers and run workloads on premise vs. the public cloud. “Organizations may need to figure out the right deployment for different workloads,” he said.

“Cost is always a factor but it isn’t the only factor,” he concluded.


Want to learn how to maximize your cloud investment? Check out our Cloud Executive Summit in beautiful Sonoma, Calif., from Dec. 6-7.