What do you call a sheep without legs?—A cloud!
—John Bishop and Jose Mourinho charity sketch.
Never mind the sheep, to think of a digital service today without a cloud-based architecture is like a table with a leg missing. To be fair, it’s not just cloud, but the layer we should think of as Digital infrastructure or the invisible part of any digital solution. This includes cloud, middleware, and security. In this chapter, let’s focus on cloud services.
What Is It?
The easiest way to think about cloud services is to think about computing power on tap. Just like electricity. You don’t have to generate your own power, or have your own generator, for which you would have to determine up front how much power you would need and therefore what kind of generator to invest in. Our homes are connected to the national grid, and we just switch on our devices, consume the energy we need, and pay for the amount used. This is exactly what cloud computing has done for information technology (IT) infrastructure and servers. It converts the entire complexity of server hardware, operating systems, configurations, and all the ancillary services into a utility model and allows you to pay as you go.
History
The idea of accessing remote computers for processing power dates back to the days of mainframes, but cloud computing, as we know it today, can be traced back to the post dot.com crash period with startups trying to build more cost-efficient models for growth. Amazon was one of the survivors who built a significant cloud environment for its own needs, and in 2006, launched a service to allow other companies to use its cloud-based Web services. Over the next few years, Amazon Web Services (AWS) matured quickly to become enterprise-grade, and coupled with the emergence of open-source software, quickly became a de facto standard for new applications being built and run at a small fraction of their precursors.
Today, the top three providers—AWS, Microsoft (Azure) and Google Cloud Platform (GCP)—are collectively known as the hyper-scalers. The global scale and performance of businesses such as Netflix, which accounts for 40 percent of the total U.S. Internet traffic has also bolstered the confidence of corporations. Nobody really questions the cloud model anymore. The default position for most large businesses is (a) multi-cloud—using more than one provider, and (b) hybrid cloud models—which means a combination of private and public cloud providers.
Meanwhile, the explosion of personal and enterprise apps and web-based consumer businesses ensured that whether or not you thought you actively considered the cloud, you’re using the cloud every time you use stream music or video, use Gmail or Google Maps, use Zoom or Teams, book a taxi on Ola, or plan your next holiday on TripAdvisor.
Why?
The most critical part of the cloud’s advantage is the flexibility. A digital project almost by definition is not likely to work to long-term projections cast in stone. The proposition, the product or app, the commercial model, may all change during the lifetime of the solution, and even over the course of the delivery of the project. Consequently, the idea that your infrastructure can grow and shrink with your needs, and services be turned on and off as required is like oxygen for the project. In the precloud era, a significant amount of cost and time would be spent on infrastructure, planning, and sizing. And that was often the largest capital commitment and risk. Now it’s a flow that you can turn on and off. And you can spend all that planning time and energy on your core business.
Even better, you can ramp down as well as up. You can set up a test bed for your alpha service, shut it down, and build a new one for Beta. This allows you to convert a lot of your capital expenditure (CapEx) to operational expenditure (OpEx), which is in essence healthier as it keeps pace with your revenue and operational scale.
It’s worth bearing in mind that the reason firms with a small operational footprint can scale globally and quickly is because of the power of the cloud, and the ability to deliver a customized service almost anywhere in the world and manage it with a small and remote team. And because new servers and entire platforms can be spun up at short notice, even programmatically, you can respond to spikes in demand almost instantly.
The XAAS Model
We spoke earlier of the conversion of infrastructure into a utility model, available on tap and on a pay as you go basis. Another way of describing this is to call it infrastructure as a service. You don’t pay to own your own infrastructure, you pay to rent it as a service. You can now rent infrastructure, software, or entire platforms using this model. This is generically known as the XaaS model. The aaS here means as a Service—and X can be applied to software (SaaS), infrastructure (IaaS), or platforms (PaaS). As a simple analogy, you might say that owning a car is like the old world of buying and owning your infrastructure, whereas a car club is a car as a service, and a taxi is a ride as a service model.
Containerization
You may come across this term while building an ecommerce application or a mobile app. Containerization is a relatively recent innovation in the application infrastructure environment. The term comes from the shipping industry where the use of standardized containers across the world allowed everybody to speak the same language and measures while shipping goods. It gave the entire industry a common unit of shipping goods. In the same way, containerization in the application space allows for the definition of a number of aspects of the server environment in which the application runs. Docker is one of the companies that was an early player in the containerization space.
Bringing Functionality to Content/Data
One of the innovative ideas from Amazon Web Services (AWS) came from the insight that while content is often very heavy and unwieldy to move around, the actual operations people want to perform on the content by themselves are much more lightweight. This led to the idea that rather than move content around in a workflow, it would be much easier to keep the content in one place and allow all the operations to be performed on the content. This could for example include adding audio descriptions, editing, encoding, and reformatting, for video content. Today, this model is being adopted by many others. For example, if you store documents on Dropbox, you can edit a PowerPoint presentation without downloading it from Dropbox, or the Microsoft Office 365 Cloud. Multiple people can access the same version and make their edits. This kind of shared workspace model is a key benefit of cloud-based environments, and it points to the cloud as a virtual shared workspace, rather than just storage.
Do You Need Your Own Cloud Strategy?
As a business user or owner, you shouldn’t usually have to create your own cloud strategy unless you’re launching a completely new digital product. Your organization or your CIO probably has a cloud strategy already with one or more hyper-scaler. Don’t be surprised, though, if there are a few AWS or Azure accounts floating around between your development teams or in the IT organization. You should probably avoid the easy route of just using a credit card and setting up another cloud account because over time, this becomes an inefficient model at scale. At least check with your IT organization about existing cloud accounts you can use.
You will definitely need to understand the cost structures and what you’re paying for, but don’t forget what you’re saving in the process. Most cloud providers have dozens of services with individual prices and easy-to-use calculators to predict the cost outflows. You should keep in mind that the cloud is no longer just about storage or hardware. Now, the entire ecosystem of your application servers, software, platforms, database, data, and the functionality of your system—all of it can live in the cloud.
Tip: Sit down with your lead developer and go through all the services you might need for your application and use the calculator to see what your needs and costs might look like under different scenarios.