I got my first computer in 1987 as a gift from my parents. I must admit that I chose an Amstrad CPC 464 because it came bundled with a few cassette tapes with games:

Figure 1.1 – Amstrad CPC 464 (author: Bill Bertram, source: https://en.wikipedia.org/wiki/Amstrad_CPC#/media/File:Amstrad_CPC464.jpg)

I used to play some of those games during my initial days with the computer, but they were unfortunately quite boring and primitive. One of them was called Animal, Vegetal, Mineral and it especially caught my attention. It was a very simple text game that tried to guess any animal, plant, or mineral that you could think of as soon as possible, by asking questions such as is it a vegetable?, is it green?, and is it a lettuce?

I have always been a very curious person and started to think how that game could work, and mentally pictured some kind of tree structure organizing all the information about potential answers so that the answers provided to the questions could help the code decide which leaf of that tree to traverse. As you can see, my mind was already a good fit for data structures, wasn’t it?

The CPC 464 came with a small book, which was one of the reasons why I became a developer (I’ll leave the other story about reverse engineering the protection scheme of a Lemmings game for another time). The title of the book in English was called Basic Reference Manual for Programmers. I loved reading at that age, indeed the worst punishment I could get was not being allowed to read at night, but this manual surprised and entertained me even more than the sci-fi classics I was used to reading because it opened the door to learning and fostered my digital creativity at the same time:

Figure 1.2 – CPC-464 Basic Reference Manual for Programmers

One step at a time, I was able to learn how to code in Basic on my own and implemented an even more primitive version of the game I mentioned earlier. But much more important than that, I loved the experience and decided at quite a young age that computers in general, and programming, in particular, would both be core parts of any professional career I would choose in the future.

Programming at that time was like writing with a pen on a blank piece of paper but using the bundled and primitive Line Editor application instead. In 1985, I just had that Basic manual to learn and get inspiration from. Forget about autocompletion, online help… and, oh my God, there was no Stack Overflow! Just take a look:

Figure 1.3 – Amstrad screen with Basic code

In the following years of my professional career, I had the chance to start working with Unix and Linux systems and started using vi and emacs as code editors for shell and Perl scripts, together with code written in C. I have always felt quite comfortable with the console, even for editing code, but I have to admit that graphical and integrated interfaces were game-changers.

A few years later, I moved to a pharma company where the presence of Windows systems was much more frequent. That meant meeting Microsoft’s Visual Studio UI for the first time, and I must admit it was a very pleasant surprise. Coding in Visual Basic felt different when compared with my good old Amstrad CPC and exploring interface design was a very interesting experience. I also started using Notepad++ for my PHP, Perl, and bash scripts. It was great to see how all these applications added new features year after year.

But I still realized I had to create everything from scratch whenever I started working on a new project, and the time required to develop an application or service was just too long.

Fortunately, nowadays, there are a lot of different integrated development environments (IDEs) to choose from, with amazing features such as code completion, code control integration, online references, and samples, which make writing code a much more enjoyable experience. I love the idea of using web-based IDEs, which in my opinion make the experience comfortable. Being able to open a browser and have access to an IDE full of options is just amazing!

Visual Studio Code (https://code.visualstudio.com/) is one example that I use quite often when I develop applications for Google Cloud. Just compare the following screenshot with the previous one from the Amstrad CPC...

Figure 1.4 – Interface of Visual Studio Code

And we are getting closer and closer to code being automatically written by just providing a summary of what the piece of code should do (https://www.forbes.com/sites/janakirammsv/2022/03/14/5-ai-tools-that-can-generate-code-to-help-programmers/).

However, what I love about developing for the cloud is how easy it is to integrate external services that provide advanced features, or how accessible architecture patterns, reference implementations, or sample code are. Now, I can put together pieces of code during a single day of work that can do much more than what I could achieve years ago working for 2 weeks.

During my first few months at Google, I worked on a solution that analyzed display ads and let advertisers know what elements worked better. Knowing whether ads with a palm tree worked better than those with a swimming pool, or whether an image of a couple with a baby got more clicks than one with a group of friends was cool.

And implementing this application was reasonably easy thanks to what Google calls the Cloud Vision API, an AI-based service able to detect objects, text, and colors on an image. Imagine how long it would take me to develop this system on my own. I will admit it: I would never have been able to do it. But now, it just takes a few minutes to integrate the API of this service with my code.

And the same happens with other key services, such as storage, messaging queues, databases, and many others that we will cover later in this book. I can say loud and clear that Google Cloud has changed the way I understand and conceive software development. Not only can I develop applications much faster and much more securely but I can also deploy them much more comfortably and make them available to any amount of users worldwide. And all of this can be done from a browser running on my laptop.

Now, I can focus my time on innovative applications and use different components, which make use of cutting-edge technologies, to develop and deploy these applications in record time. And this is what Google Cloud can do for you, too.

So, long story short, after spending endless days fixing the effects of bad initial designs and upgrading hardware, operating systems, and applications, I realized that it would be much better if I could design and build resilient and distributed applications while reusing cutting-edge components and services, which would scale great and deploy faster, and try to decouple them from the hardware and the operating system.

But when we talk about developing, it’s not just writing code but also getting to identify what problem or challenge needs to be solved and decide how we can provide a solution that works. And doing all of this properly is the real challenge.

In my case, I was lucky because, just before joining Google, my employer happened to start exploring ways to make projects shorter and more successful. That’s how I became familiarized and started to use project management, Agile, DevOps, and, once I joined Google, Site Reliability Engineering (SRE) practices.

Let’s discuss what these are and how they can help developers.

Organizations put a lot of effort into minimizing the waste of time and money in projects since both are frequently scarce. Being able to anticipate bottlenecks and blockers can help reduce the chances for a project to fail.

And it is here that project managers become the key players. Among other tasks, they are responsible for identifying stakeholders, dividing the work into tasks, assigning times to each activity, and following up to ensure that everything is completed on time.

Traditional project management used the so-called waterfall methodology, which divides a project into different steps that are completed in sequential order one after another: requirement gathering, design, implementation, testing, and maintenance.

However, there can be projects that may run for much longer than planned due to different reasons – for example, wrong or incomplete initial assessments leading to undetected dependencies, or never-ending tasks that block others.

Also, projects managed using waterfall methodologies are more rigid in terms of features. As these are defined in the initial phases, any changes due to unexpected reasons, such as a feature not being needed anymore or becoming obsolete, could derail the project.

Project management has evolved and one of the most common practices to reduce the risk of long delays is to split the project into different phases of incremental complexity, also known as sprints, while following an iterative approach instead of a linear one. These practices were introduced in more recent methodologies, such as Agile, which aim to speed up the progress of projects and offer tangible results as soon as possible.

In Agile and similar methodologies, a Minimum Viable Product (MVP) can be provided after completing just one or a few of the initial code sprints; then, the team will work on improving it using an iterative approach that adds new features and capabilities. It fixes any found bugs in each new sprint until the project meets all the requirements and is then considered to be finished.

The following diagram summarizes the different phases for each sprint:

Figure 1.5 – Agile development phases

Agile is a project management methodology aimed at getting an MVP ready earlier, but it needs a compatible process on the development side to ensure agility. And here is where DevOps comes to the rescue.

DevOps is a set of practices that aims to increase the software delivery velocity, improve service reliability, and build shared ownership among software stakeholders. Many organizations use DevOps to complement Agile project management methodologies and reduce the lead time – that is, how long it takes for a team to go from committing code to having code successfully deployed and running in production:

Figure 1.6 – DevOps development cycle (source: https://nub8.net/wp-content/uploads/2019/12/Nub8-What-is-Devops-1-min.png)

By implementing DevOps, you can improve a lot of your development metrics by increasing the speed of your deployments, reducing the number of errors in these deployments, and building security from the start.

These methodologies are very interesting for developers, but agility can only be achieved if the underlying infrastructure components are also compatible with fast deployments. For example, running short and fast sprints will not make sense at all in a platform where virtual machines are provided no earlier than 3 days after being requested and databases after 5 (and I have seen that happen, I promise).

An environment that can help you speed up all your processes is the best option, not only for developers but for everyone involved in IT projects. As we’ll see shortly, the cloud is an extremely good option if you use Agile methodologies, are a big fan of code sprints, or want to implement DevOps in your organization.

And if DevOps helps automate deployments, SRE can also help in a later stage by automating all the manual tasks required to keep your environments up and running, such as those included as part of change management or incident response processes. And guess what – the cloud is a great place to implement SRE practices, too! To learn more about SRE, visit https://sre.google/.

If you are new to concepts such as Agile or DevOps, you may still be wasting a lot of your precious time as a developer doing the wrong kind of things. You should be spending most of your hours on innovating, thus contributing to the Digital Transformation of your team and the whole organization. We’ll use the next section to explain what Digital Transformation means and why it is really important and should be appropriately prioritized if it hasn’t been already.

I can imagine that many of you, while reading the first part of this introductory chapter, will have remembered your very own unpleasant experiences of working with infrastructure, applications, and architectures that started to grow and run out of resources due to a limited physical or virtual on-premises environment, a monolithic or overcomplicated initial design that made the application or service die of success after growing much more than expected and that you had to fix for good, or data split among so many databases in the organization that a minor update in the schema of a supposedly rarely used table broke most of the corporate applications.

The situations I just pictured are quite common among organizations that are still using an important amount of their IT time to decide where their infrastructure should run. And that’s probably because they haven’t completed their Digital Transformation yet. Even if you work for a start-up in its first stages, you may still be asking yourself these kinds of questions today. If that is the case, you should embrace the practices of digital transformation starting today.

The reason is that all these sadly common situations are incompatible with innovation. And IT professionals in organizations where innovation is constantly postponed because there are other higher priorities will become either outdated or burnt out, if not both, over time. If we combine this golden jail scenario with the burden of system and infrastructure migrations, there is a lot of precious time wasted on tasks that developers and engineers hate, and that don’t add any value to the organization.

Let’s say it loud and clear: if you want to innovate and if you want to be disruptive, you should focus your efforts on transforming or creating an organization where everyone can drive innovation. Otherwise, you will be wasting precious time and resources focusing on the wrong tasks.

Rob Enslin, former President of Global Customer Operations for Google Cloud, mentioned a few areas to focus on during a digital transformation process in a blog post from the Google Cloud website: https://cloud.google.com/blog/topics/inside-google-cloud/innovation-in-the-era-of-the-transformation-cloud. This list is, in my opinion, a very good summary of four of the main pillars of digital transformation, where organizations should put their efforts to free time and resources and be able to innovate more.

Let’s comment on each of these pillars:

• Accelerate the transformation, while also maintaining the freedom to adapt to market needs. This is a very important point because while the digital transformation should happen in a reasonable amount of time, the process itself needs to be flexible too; otherwise, it may fail miserably if either the market or any other important external variable suddenly changes without prior notice. For example, during the pandemic, many companies were forced to speed up their digital transformation, and those who were ready to provide remote working capabilities for their employees earlier suffered less from the effects of the lack of productivity during those months that all of us had to spend working from home.
• Make every employee, from data scientists to sales associates, smarter with real-time data to make the best decisions. First-party data is power; however, it is often split into silos across an organization. A digital transformation should break down these silos by centralizing, deduplicating, and consolidating all data sources so that all the information is available to all members of the organization together with real-time insights that each department can use to make their own informed strategical decisions.
• Bring people together and enable them to communicate, collaborate, and share, even when they cannot meet in person. After the pandemic, it’s even more clear that physical distance should not be a stopper, and all cultural elements of the organization should be replicable for people working remotely too so that people can also collaborate and share comfortably when they are far away from each other. Consider this as flexibility seen from a very specific angle.
• Protect everything that matters to your organization: your people, your customers, your data, your customer’s data, and each transaction you undertake. Security is more important than ever, especially now that companies are using the power of technology to provide better services, and it should be a key element in any modern company transformation plan. Your data is your treasure and, together with your intellectual property, it might be what differentiates your organization from the competition. But it is also your responsibility to keep all your data safe, even more so when it probably contains personal and private information about your customers.

Rob summarizes these four pillars into their corresponding objectives: application and infrastructure modernization, data democratization, people connections, and trusted transactions. Any organization able to meet these objectives will have much more time and resources to dedicate to innovation.

If you read the previous paragraph carefully, you will realize that we developers are the key players in each of the four pillars of Digital Transformation, one way or another. During the digital transformation of an organization, developers will be working hand in hand with engineers on application and infrastructure modernization, which should be achieved by simplifying monolithic architectures by splitting them into elastic microservices. These apps and services will be using data as an input, and probably also generating data and insights as an output in many of the cases, so they will benefit from both the data centralization and the democratization mentioned earlier, and code should become simpler once data is easier to access.

And being connected to the rest of the team will also be important to make sure that our code meets everyone’s needs. If we work using sprints, we need to be aligned with the rest of the team, even if each of us is located in a different office, country, or even continent. Finally, security is the key to ensuring that our apps and services are safe to be used and that our customers trust us more than ever.

Designing a Digital Transformation plan is not easy, and that’s why there are a lot of companies working to help others succeed on their transformation journey. Some companies can help you design and execute the plan, but many others have created platforms that can make things much easier.

Some years ago, tech giants had the idea of abstracting the infrastructure up to the point that the customer wanted, letting organizations focus on what they love to do: architect, write, and run modern applications, centralize their data, make the most out of it, and get people connected, all of it in a secured platform. And guess what – Google Cloud is one of them.

There are a few reasons why I would recommend developers run their code on the cloud.

First of all, let me say once again: if you are spending too much time setting up servers, installing operating systems, deploying patches, and performing migrations, then you simply deserve better. I’ve been there and I can’t put into words how happy I felt after I left it behind. I used to dedicate 20% of my time (and much longer on specific occasions) to maintaining, troubleshooting, and supporting the operating system, applications, and database. Since I joined Google, I can use that extra time to learn, brainstorm innovative solutions with my team, or write better code. I also think that code is poetry (https://www.smashingmagazine.com/2010/05/the-poetics-of-coding/), so IMHO, inspiration arrives better when we have more time and less pressure.

Besides, most cloud-based services offer customizable infrastructure components, or at least different sizes, so you can still have a reasonable degree of control over where your code runs. In summary, running code on the cloud will provide you with more options and better performance and will allow you to focus your time on coding, not on other distracting tasks.

Also, a cloud provider has many data centers in different locations across the world. If the start-up or organization you work for is planning to grow and have customers in more than one market at some point, a single server or even a few servers in a single location may not be enough to offer decent-quality service. This is becoming more and more important as real-time services become predominant and latency has to remain low.

If you can anticipate that you may suffer a potential scaling or latency problem in the future, already being in the cloud can make things much easier if you need to replicate your architecture in another continent for local users. Having infrastructure located closer to users can also help you meet legal requirements, as some countries require data or processes to be located in-country.

And speaking about scaling, the pay-per-use model is reasonably compatible with organizations growing because you will use more resources as you make more business and generate more revenue. Besides, most cloud providers will offer bigger discounts as you increase your usage. And if you have very particular needs, you can use huge clusters for a few minutes or hours and pay a very reasonable price. Indeed, you can have thousands of servers at your command at a very affordable price, something that would be prohibitive in an on-premises data center.

If your infrastructure is affected by traffic peaks, the cloud is also your place to go. If you have a lot of visitors on Sundays, your website crashes during Black Friday, or your app is usually down during the Christmas holiday season because of seasonality peaks, you may have decided not to increase the resources available for your application or website because, during most of the year, it can handle the average demand. With a cloud provider, you can scale up your application automatically when there is a peak and you can do the contrary too – that is, you can scale it down while your customers are sleeping so that you can reduce your costs. You can also schedule some tasks to run when the data center has less workload and save more money. We will discuss all of these opportunities later in this book.

What if you want to implement Agile and DevOps practices? Cloud providers have very fast provisioning times, so you can deploy complex applications and the infrastructure associated with them, as it is no longer static, in a matter of minutes. And that makes a huge difference, which allows you to use that extra time for better testing or even to do more iterations, which in the end will translate into better code.

And regarding the everyday life of a developer, if you are worried because you may not be able to keep on using your favorite IDE or fear that latency while writing code could be a problem, or that processes might be more complicated, just give it a try – you will be delighted. Hosting your repository in the cloud should be easy and you will not notice the difference. And you can connect from anywhere, even while commuting back home if you realize that you forgot to submit a very important CL before leaving the office.

I hope that you have been convinced that running your code on the cloud is a great idea. Now, let me show you why I think that Google Cloud is the best cloud provider to do so.