Bare-metal or dedicated servers are fantastic when you need speed and power on a single operating system (OS). You have one OS on one physical machine. But, it is possible to split up one physical machine into multiple virtual computers, that act like physical computers and run their own operating systems. These computers are called virtual machines, or VMs.
To do this, you can install a layer of software called a hypervisor. This process is called virtualization. The hypervisor acts as the layer allocating hardware resources between the bare metal’s OS and the OS of the virtual machines. That means that even if you are not the single tenant of a bare-metal server, you will still experience your own VM unit as a simulation of a dedicated server.
There are two types of hypervisors. The purpose of both types is the same; to provide a minimal operating system for virtual machines. However, Type 1 runs directly on the bare metal, while Type 2 runs on top of the pre-existing OS.
A bare metal hypervisor is also called a Type 1 or a Native hypervisor and is the type that is deployed the most. A bare metal hypervisors is installed directly on the hardware, often in place of the operating system. It has direct access to the machine’s resources, as well as running the guest OSes and their applications.
Bare metal hypervisors are generally extremely secure because they avoid the flaws and vulnerabilities that are associated with attack-prone OSes.
Type 1 hypervisors perform much better than Type 2 as the absence of a middle layer means the resources of the hardware can be accessed more directly.
Not being dependent on an underlying OS makes a VM with a Type 1 hypervisor much easier to scale.
Type 2 hypervisors are also known as client or hosted hypervisors. A Type 2 hypervisor runs on top of the operating system already installed on the machine. They rely on the host machine’s OS to manage the requests to the hardware resources like the CPU, memory, and storage. This is why they are called “hosted” hypervisors because they are “hosted” by the original master OS.
Type 2 hypervisors tend to experience higher latency because the requests to the hardware from the hypervisor have to go through the extra software layer of the operating system.
Using a shared space often increases security risks, and threatens the performance of a VM due to the noisy neighbor effect.
Scalability is limited as a VM with a Type 2 hypervisor is reliant on the operating system of the machine.
While bare metal may still come up as best for high performance and security, hypervisors work very effectively for dynamic workloads, scalable or smaller use-cases such as edge computing.
Bare metal hypervisors (type 1) are well suited to the large-scale requirements of enterprises and data centers. Such organizations will have many servers, and many use cases. Using hypervisors enables physical resources to be used more efficiently.
Type 2 hypervisors are best suited to circumstances where performance or security are not an issue. They are especially best for personal or small deployments. Hosted hypervisors are also used for development and testing on your own machine, by spinning up some VMs so you can directly monitor your work and results.
The hypervisor is software layer that separates a computer’s operating software from it’s physical hardware and translates requests between physical and virtual components. In doing so, the hypervisor gives the physical machine the ability to run many virtual units. Hypervisors come in two types, Type 1 and Type 2.