Google Compute Engine

At the lowest level, we have IaaS, which provides users access to fundamental components such as virtual machines, networking, storage, as well as tools and services for managing and interacting with these components. For Google Cloud, IaaS primarily takes the form of Compute Engine, and is deeply integrated with Google's networking solutions such as load balancers and virtual private clouds, often collectively referred to as Network as a service (NaaS).

By exposing direct control over the underlying infrastructure, IaaS provides an unparalleled level of control and flexibility for building cloud solutions. Although most IaaS services are generally considered low-level, in reality these services provide a fairly high level of abstraction. For example, when users request Compute Engine VMs to run a given image with a set amount of vCPUs and RAM, and Google Cloud abstracts away the process of securing these resources and creating virtual machines running the given image. This level of abstraction over the underlying infrastructure maintains most of the power of direct infrastructure control, while providing a rich control plane that speeds up operational efforts and facilitates practices such as Infrastructure as code (IaC).

While having complete control over the underlying infrastructure of a solution brings an incredible amount of power and flexibility, this control can represent a significant cost of developer and operations effort. For many solutions, an ideal environment would abstract away higher levels of the underlying infrastructure, allowing teams to put more effort into solving business problems.