Understanding Hypervisors: The Core of IaaS Virtualization Explained

This is the second installment of the HULK virtualization team’s series, following an overview of cloud computing history and now focusing on the underlying IaaS virtualization layer, using everyday analogies to make advanced technology easy to understand.

In IaaS, users receive a virtual server—e.g., a 4‑core CPU with 8 GB RAM—without worrying about the physical hardware or its location. That server is actually a virtual machine (VM) created by software.

The article compares this to the fictional "split‑body" skill, where a single entity creates multiple virtual copies; similarly, a single physical server can host many isolated VMs.

The key enabling component is the hypervisor, which manages these virtual copies. Hypervisors come in two categories:

Bare‑metal hypervisors run directly on hardware (e.g., VMware ESXi, open‑source Xen).

Hosted hypervisors run as a module on an existing OS (e.g., KVM on Linux or Windows).

Understanding the history of Linux virtualization clarifies why Xen’s code could not be merged into the Linux kernel, while KVM, built into the kernel, quickly gained strong support in OpenStack and is the hypervisor used by the author’s team.

Both hypervisors have pros and cons, but the article emphasizes two primary benefits of virtualization:

Cost: higher resource utilization.

Management: greater flexibility, such as quickly reallocating a physical server’s capacity among team members.

For example, AWS started by sharing idle resources and has become a major revenue source for Amazon.

The next article will provide a hands‑on experience of this "split‑body" technique.