Mainframe computers, like the IBM one shown above, were the first digital computers that started appearing in big companies during the 1950s. They were room-sized behemoths of circuitry and steel primarily used to generate balance sheets and income statements. And while IBM still sells a small number of mainframe computers today, they only enjoyed mainstream popularity until the late 1980s.
While early mainframes were only accessed by the people who managed it, most mainframes of the 1970s and 1980s ran programs that were accessed by remote users via a dumb terminal. Dumb terminals were not standalone computers; they had a display and keyboard, but merely allowed you to interact with a central mainframe computer across a network connection.
To allow for many user connections, each mainframe computer consisted of multiple separate processors, each designed for a specific task: compute, disk storage, networking, and so on. As you added more users and dumb terminals, you could add more processors to the mainframe to accommodate them.
Unfortunately, mainframe hardware was too expensive in the 1970s and 1980s to ever make its way into small- and medium-sized organizations. Luckily, the personal computer (PC) revolution was gaining momentum in the 1980s, and the IBM PC (alongside its numerous compatible clones) quickly became the standard for business computing as it was affordable for even very small organizations.
The first IBM PC (model 5150, shown here) was powerful enough to run programs such as spreadsheets and word processing applications with ease. Plus, you could connect it to a computer network and access other computers (called servers) to store files, access databases, or even print to a shared printer. Like mainframes, PCs were scalable; small organizations may only need a few of them to keep track of business transactions and inventory, while larger organizations could have hundreds or more.
Compared to mainframes, PCs had a very simple design. They only contained a single processor (called a central processing unit, or CPU), with all other functionality (sound, graphics, networking, keyboard, mouse and disk storage) was provided by additional circuit boards.
By the 1990s, most organizations were performing critical business functions on PCs, and these PCs were connected to the Internet to allow for Web browsing and email. PCs got more and more powerful, but largely retained the same single processor design until the late 2000s, when it made more sense to incorporate multiple processors on a single chip (each of these processors is called a processor core). This was vital for smartphones, where the number of components had to be kept to a minimum and use as little battery power as possible.
A typical processor in your smartphone today has multiple separate processor cores dedicated to running your programs (apps), as well as performing other tasks such as generating graphics, accessing storage, providing security, accessing wireless networks, and so on. Moreover, most modern smartphone processors contain multiple processor cores dedicated to just running apps (my phone has 8 of them). These additional processor cores allow the smartphone to run apps faster.
In other words, the processor in your smartphone is much like a mainframe computer; it contains multiple processor cores, including many that are dedicated to a specific task.
The same is true for the processor in your desktop or laptop PC this past decade. However, the processor cores in a PC often run much faster and use more power than those in your smartphone.
This past decade has also seen a shift towards computing on powerful servers with many processor cores, hosted in data centers available via the Internet. The worldwide collection of these servers is called the cloud, and leveraging them is referred to as cloud computing. Today, cloud computing is ubiquitous for both personal and business use. Cloud-based services like Microsoft 365 (formerly Office 365), Zoom, DocuSign, Zendesk, SAP, Dropbox and Slack are used within most organizations today, and millions of other cloud services provide the backend functionality for online shopping, social media, video streaming, and Internet-enabled devices such as thermostats, fridges, home automation devices and virtual assistants (e.g. Amazon Alexa).
To access cloud services, all you need is a smartphone or PC with Internet access, as well as a Web browser or local app. Nearly all of the computing is performed on the cloud server, while the Web browser or local app delivers the results. Consequently, you don’t need a powerful computer to use cloud services; an inexpensive Chromebook or budget smartphone will give you the same experience as a high-end PC. In many ways, the Chromebook and budget smartphone are dumb terminals that users use to access cloud services across the Internet, while the cloud servers are mainframes that provide the central compute power.
The mainframe computing paradigm is alive and well today, and the tech industry shows no signs of moving away from it in the future. CPU designers are continuing to add more specialized processor cores with each revision, and cloud computing is continuing to shift heavy processing away from users and towards servers in data centers.