Cloud computing history part 1: the origins of the cloud

The origins of cloud computing go back further than you might think. Learn about the forebears of the modern cloud in our cloud computing history guide.

How long do you reckon the cloud has been "a thing"?

If you're a casual computer user, you might say 10 years or so. After all, the early 2010s saw the launch of many public-facing cloud storage services, such as Apple's iCloud and Google Drive. 

We'd guess this is when the concept of the cloud lodged itself in the public consciousness – or at least became acceptable to discuss at dinner parties. 

If you're a techy type – a sysadmin, say – you might plump for the 2000s when the first large-scale public clouds launched. Or the late 1990s, when we felt the first murmurs of a tectonic event we'd later call "SaaS".

Good guesses. But it might surprise you to learn that the origins of the cloud go back much further.

How much further? Ooh, say… 30 to 40 years?

1959: the beginnings of virtualisation

Virtualisation is one of the pillars of cloud computing today. It's what allows us to share memory, compute and networking resources between thousands of users, all around the world.

To put it another way, without virtualisation software like the VMware SDDC , a server would just be a server – a discrete piece of hardware that could only do one job.

Virtualisation is not new. The world got its first taste of the concept way back in the 1950s when some very clever computer engineers at the University of Manchester invented virtual memory.

After successful prototype demonstrations in 1959, the technology was rolled into the Atlas Supervisor. This program was, among other things, responsible for managing virtual memory for the Atlas Computer, one of the world's first supercomputers.

The concept was later expanded from memory to other computing functions, with the first full-fledged virtual machines arriving in the 1960s.

Before that could happen, though, we had to wait for some geniuses to invent a little concept called time sharing.

1961: the first time-sharing operating system

OK. Before we go any further, let's get one thing straight: time sharing is not the cloud.

If anything, time-sharing systems were more like early multi-user local networks. They allowed multiple programmers to access a mainframe computer at the same time.

Before time-sharing, the predominant means of accessing compute resources was through batch processing. Programmers would write code "offline" using punch cards and then submit the cards to the operations team, who would schedule a time for the program to run.

Slow? Not half. An individual programmer could expect to wait days until they saw the results of their program.

Time-sharing sped up this process by sharing compute resources between users. These users could enjoy direct access to the mainframe without causing expensive periods of downtime.

So, yes. Time-sharing is not the cloud. But by pioneering the sharing of compute power, it stands as an important milestone on the journey to modern cloud computing.

1969: ARPANET pioneers wide-area networking

We can't possibly talk about the history of cloud computing without mentioning ARPANET, the granddaddy of the modern internet.

ARPANET was like the internet in miniature, without fancy-smanschy features like websites, graphics or bingeable, streamable TV shows.

It was hardly "miniature" for its time, though. As the world's first wide-area packet-switched network, it connected users over previously inconceivable distances. By 1957, the original four-node network had expanded to become a truly international WAN, with nodes popping up in the UK and Norway, as well as in its native US.

ARPANET was special. But while it's easy to get excited about its scope and influence, we'd argue that its greatest contribution to computing was something a little more mundane: the TCP/IP protocol.

See, as the world cottoned on to ARPANET's potential, several copycat networks sprung up around the world. And each of these networks spoke a different language – there was no way to connect your CYCLADES node in France, say, to the ARPANET at large.

TCP/IP changed all that. As a general protocol, it acted like a skeleton key for WANS, unlocking the  doors that separated one wide-area network from the next.

As TCP/IP was adopted around the world, the boundaries between networks became increasingly blurry. Eventually, it made sense to lump all the networks together under one catchy name: the internet.

TCP/IP is still used today. It forms the building blocks of the modern internet and – by extension – cloud computing.

1958: SAGE and the early cloud data centre

Have we just travelled back in time? Yes, but there's a good reason for it – bear with us.

ARPANET was very, very important in the evolution of cloud computing. But, fundamentally speaking, it existed to connect users with users.

Cloud computing is all about connecting users to ultra-powerful hubs of juicy computing horsepower. Or, as most people call them, data centres.

To complete our cloud-shaped jigsaw puzzle, we need to look back a little further to the precursors of the modern cloud data centre.

One such precursor is SAGE (the Semi-Automatic Ground Environment), which launched in 1958.

SAGE was a military C3 (command, control and coordination) system. It gathered information from radar sites across the United States and combined it into a single radar image that covered a wide area.

SAGE wasn't a centralised data centre – it operated from several sites ("Direction Centers") built on airbases around the country. However, in its function as a data aggregator and processor, it had many similarities to what we'd call a private cloud data centre today.

It didn't look much like a modern data centre, though. It was powered by some of the largest and most expensive computers ever built, each of which demanded an astonishing 22,000 square feet of floor space.

These computers could manage 75,000 instructions per second. To put that into perspective, the processor on an iPhone 6 can process 3.36 billion instructions per second. That makes the pocket-sized iPhone 44,800 times faster than the massive hunks of computers that powered SAGE.

The trend for smaller, faster processors would be a key driver in the evolution of cloud technology. But you'll have to wait until next time to hear about that. Stay tuned.

Ascend Cloud Solutions is a cloud consultancy and migration firm founded by ex-VMware engineers. Learn more about our cloud consultancy services today.