Intro to Virtualization and the Cloud

Virtualization allows for the creation of “virtual” computers and devices on one server. It is an effective strategy for partitioning and managing compute, storage, and network resources. Virtualization is not a new concept, but its popularity has waxed and waned over the decades.

Back in the days of mainframes, time-sharing was invented to allow multiple users to run code on one mainframe, but it had its limitations, so virtualization was created¹. This allowed each user to seemingly have their own mainframe, and was a great boon to administrators who wanted to keep the hardware at the highest possible effective utilization. With the rise of personal computers, virtualization took a backseat to the new user-friendly desktop machines—why seemingly have your own machine when you can actually have one? As operating systems advanced, the original problem which time-sharing and virtualization solved became more of a OS implementation detail of how to best manage multiple processes.

Fast forwarding past various other operating system advances, the development of the Internet and its protocols, and the general exponential increase in information technology (which I feel lucky to have observed during my life), and modern cloud providers find themselves in much the same position as the early managers and developers of mainframe computers. The central problem is still how to most effectively keep their machines busy and allow users—now paying customers—access to compute, storage, and networking resources. By virtualizing these aspects of computing, the cloud provider takes care of the underlying infrastructure and allows customers to provision whatever resources they need.

While customers can provision storage and compute capabilities in the cloud, and run a growing variety of services, all these components have to talk to each other. In a traditional network, physical devices play a large part in the creation of that network, for example: switches connecting devices and routers managing network traffic. By using software-defined networking (SDN), these physical devices can be abstracted away and seemingly created on demand. The benefits are enormous. By defining compute, storage, and network capabilities programmatically through infrastructure-as-code (IaC), we can roll back to a last known good configuration if any recent experimental changes prove untenable.

Since the advent of time-sharing, one of the more complex issues in managing computer resources was maximizing the ability for multiple users or processes to take full advantage of the hardware². Virtualization of compute, storage, and networking in the cloud allows organizations to quickly productionize IT infrastructure, respond to change in demand for specific resources, and avoid many of the limitations of traditional on-prem infrastructure.