After 17 years of cloud computing, what lessons have we learned?

Cloud computing has become a vital part of the corporate IT landscape. Not a controversial statement, but head back 17 years to the launch of Amazon’s S3 and the world was very different. IT departments protected their on-premise servers as if they were Gollum’s ring, with nothing allowed to interfere with their precious. At best, companies ran a few trial projects, often starting with storage.

It seems inconceivable now that there could be such negative attitudes to cloud computing. Particularly at a time when, according to research from O’Reilly, more than 90% of enterprises are running cloud applications.

So, something momentous has happened in the past decade and a half. What we want to explore here is the lessons learned… and the lessons we still need to learn.

Security issues haven’t gone away

IT’s main concern in the early days was that cloud was insecure. Hardly surprising. In a world where security was the main concern facing IT directors, there was natural unease about handing company data to a third party. In particular, people worried about multi-tenant operations and whether their data would be safe.

Even by 2015, a survey from the Cloud Industry Forum found that 70% of IT decision makers stated that data security was one of their biggest concerns when deciding whether to move to the cloud.

During the time that cloud has emerged as a germ of an idea to an essential part of the IT armoury, security as an issue has not gone away – it’s just that the concerns are slightly different. Back then, IT departments worried if employees could be trusted enough and whether their processes were secure.

Now, our concerns centre range from confidential computing to how to balance seamless user access (across all devices) with data protection concerns. One thing is certain: security worries will never go away.

Not all cloud providers are the same

It’s easy to think of cloud as a generic term. If one provider doesn’t satisfy, then swap for another! But that’s not really the case. Each provider has its own strengths, its own range of products that aren’t provided by rivals. For example, Google is particularly hot on data analytics and has carved a niche in that area.

There are geographic differences too. Alibaba is clearly very strong in Asia but hasn’t had a big impact – so far – in Europe. But this doesn’t have to be about continents, it could simply be about where the nearest cloud datacentres are, either for latency reasons or because of compliance issues. Read our separate article about data sovereignty.

That’s all before we take into account that some cloud providers take aim at particular vertical sectors. Some providers have developed offerings targeted at specific industries that take into account their needs. For example, businesses in the financial sector must put an emphasis on compliance.

Cloud is not always the cheapest option

The first customers for cloud were quite clear about the motivation for such a move: cost. True, not every company considered this the main factor, but a big majority did. This manifested in two ways: the pay-for-use aspect of cloud, and that it could be classed as an operational expense rather than a capital cost.

The reality was different. Many cloud customers found that their costs were rocketing and their budgets hit. According to a 2022 Gartner report, cloud costs are, on average, two or three times higher than expected.

There are several reasons for the discrepancy. First, the unavoidable associated cost of moving to cloud. For example, adding an alternative disaster recovery plan, more complex management and monitoring, training for a new set of skills.

Then there’s the question of ease of use. Because it’s so simple to fire up a cloud instance to handle a new workload, companies would just do it. But very often, these weren’t shut down when completed.

The spectre of shadow IT once again rises. Often, users weren’t telling IT about the services they ordered and buying them at unfavourable rates, sometimes duplicating resources. It also proved to be too easy to over-provision and introduce more compute resources than were needed.

CIOs are now getting a better handle on costs. However, as plans for cloud are becoming more ambitious, there are still going to be some anxious CFOs out there.

Don’t ignore skills

Cloud computing has completely changed the way that businesses operated. IT staff who have been used to working one particular way, using specific technologies, had to get used to a new way of working. Companies realised that either employees needed to overhaul their skillsets or they needed to find new employees with the appropriate skills.

Unfortunately, every other company had the same idea, meaning that individuals with requisite skills became in high demand. All of which added to the costs mentioned above.

As companies look to do more complex and advanced things with cloud, and deploy newer technologies, the need for a broad range of cloud skills becomes more pressing.

According to a 2022 report from IDC, businesses are struggling to complete cloud transformation projects due to the lack of cloud professionals.

Indeed, about a third of respondents to the survey said this shortage had prevented them from launching services. The report highlighted a lack of expertise in three areas: multicloud, system development and governance.

Cloud doesn’t stand still

We’ve come a long way since Amazon launched its S3 service, on 14 March 2006, and created a cloud industry. Since then, Amazon has kept itself ahead of the game and many of the innovations that have transformed the cloud have emerged from the company.

It would be foolish, however, to think that there’s one version of cloud and that’s unchanging. 2023’s cloud is markedly different from the first offerings from Amazon and its successors.

There’s the emergence of a whole generation of cloud native offerings, with a tighter integration of software with the cloud. For example, Kubernetes has transformed the way that enterprises build their infrastructures.

The growing use of AI and increased automation will change the way that cloud is monitored and managed. We’ve seen the rise of serverless computing, a concept that would have been alien to the first cloud mover. Then there’s the emergence of SASE (secure access of server edge), the integration of security with cloud.

All of these will require a new set of skills and a need for closer management. But these new technologies show that cloud is continually evolving and will do so for some time yet.

Time doesn’t stand still, and nor does cloud computing.

Maxwell Cooter
Maxwell Cooter

Although Max trained to be a programmer, he quickly found his vocation in journalism. He was the founder editor of Cloud Pro, the UK's first dedicated cloud publication and has written for dozens of titles, including The Guardian and The Daily Telegraph. At TechFinitive he writes about cloud computing and data.