Cloud Computing: the invisible orchestra of modern software

Social media, video streaming, music, live events, video games, artificial intelligence, e-commerce, public administration: every day we live in the cloud without even realizing it. All you need is a laptop or smartphone and a good connection, and everything is at your fingertips.

But it is worth asking:

How do giants such as Meta, Alphabet, Microsoft, Amazon, and OpenAI manage to guarantee continuous access to billions of people?

The answer is: Cloud Computing.

The cloud orchestra: distributed systems

To truly understand the cloud, we need to start with a more general concept: distributed systems.
A distributed system is simply a set of machines, often scattered around the globe, connected to each other in a network. To put it simply, this network of machines works together in parallel, behaving like a single huge computer. They are, in fact, a distributed (many) system (one).

A distributed system is characterized by:

• scalability: the ability to add new resources (memory, computational capacity) without having to redesign the entire system;
• redundancy: the same functions are performed by multiple nodes in the system, so that if one fails, the others keep the service online;
• fault tolerance: ability to continue operating even in the presence of hardware or software errors;
• transparency: the user does not perceive complexity, but only the result.

Distributed systems are nothing more than an orchestra performing software.

The symphony of a click

As we mentioned earlier, behind every click there is a whole world that we cannot see. Without our knowledge, a network of servers runs software, exchanges data, and returns results. A digital symphony.

In this musical parallelism, we have focused on certain roles:

• the orchestra, i.e. distributed systems
• the partition, i.e. the software executed
• the listener, all of us on the other side of the screen.

Three key figures are still missing: the conductor, the person who enhances the music from behind the scenes, and the system that makes everything accessible. Let’s reveal them together!

The conductors are the major cloud providers — AWS, Azure, Google Cloud, OCI, to name a few. They set the timing, priorities, and flows. If one node goes down, another one kicks in. The performance must never stop.

The technicians behind the scenes are the companies that build the services we use: Spotify, Netflix, Prime Video, OpenAI.

Finally, the protagonist of this article: the cloud.

The cloud is the sound system for the entire concert: it amplifies, distributes, and makes it accessible. It is what allows technicians to create extraordinary services, providers to scale and conduct the orchestra, and us to enjoy it with a click.

Orchestra conductors: major providers and their technologies

As we have explained, every digital service we use relies on complex infrastructure orchestrated by cloud computing giants. While all these providers guarantee availability and scalability, they also seek to differentiate themselves by offering unique technologies that influence how software is designed, distributed, and managed.

Musical instruments: service models

Let’s continue our musical journey by focusing on the instruments made available by the cloud. Just as instruments allow us to choose the sound we want to create with our hands or breath, service models allow us to choose the level of abstraction that best suits our needs, from ready-to-use software to total control over virtual hardware.

SaaS – Software as a Service

This is undoubtedly the most familiar face of the cloud: ready-to-use software, accessible via a browser or app, with no installation or manual updates required. Examples include Gmail, Microsoft 365, Salesforce, Figma, Spotify, and Duolingo.

FaaS – Functions as a Service

We go behind the scenes and into the technical side of things. Behind every click, every automation, every invisible trigger, there is often a serverless function. With FaaS, software developers take a leap towards purely logical abstraction: they only write the logic and the cloud takes care of everything else. The main examples are AWS Lambda, Azure Functions, Google Cloud Functions, and Oracle Functions. They are ideal for implementing microservices, lightweight APIs, automations, and asynchronous processing. The main advantages? Low cost (on-demand), automatic scalability, and zero maintenance.

PaaS – Platform as a Service

When logic isn’t enough and it’s time to get your hands dirty, cloud platforms come to the rescue. They are the perfect laboratory for developers: environments ready for writing, testing, and distributing code, without worrying about the underlying infrastructure. The main examples are Heroku, Google App Engine, Azure App Service, and Oracle Application Container Cloud. The main advantage is being able to put the concepts of continuous development and continuous integration (CD/CI) into practice.

IaaS – Infrastructure as a Service

At the lowest level, we enter the cloud’s engine room. Virtual servers, storage, network. At this level, the user has total control over the operating environment. Here are some examples: AWS EC2, Azure Virtual Machines, Google Compute Engine, Oracle Bare Metal.

The ticket: sustainability and costs

Behind the convenience of the cloud lies a physical reality made up of data centers, energy consumption, and infrastructure choices that have a direct and significant impact on our environment and our economy.

In fact, every click, every algorithm, and every AI model generates a computational load that must be powered and cooled. Based on data from various sources, it is estimated that:

• Hyperscale data centers consume between 20 and 50 MW per year, enough to power up to 37,000 homes.
• Refrigeration can account for up to 40% of the total energy consumption of an AI center.
• The water used for cooling can exceed 1.7 million liters per day.

Generative AI has only accelerated the cloud energy crisis. Just consider that training and inferring language models (such as ChatGPT) requires very high-performance GPUs, often in clusters of hundreds of units, and that according to the IEA (International Energy Agency), global data center consumption will rise from 460 TWh in 2022 to over 1,000 TWh in 2026 (here).

History snippets

As we mentioned, training the artificial intelligence that we all use requires enormous computing power. NVIDIA GPUs reign supreme in this field, serving as the true beating heart of AI-ready server farms.

NVIDIA was founded in 1993 by Jen-Hsun Huang, Chris Malachowsky, and Curtis Priem. Their goal was to create graphics cards for video games, and the key moment was the launch in 1999 of the GeForce 256, considered the world's first GPU.

In 2000, NVIDIA acquired its rival 3dfx and signed an agreement with Microsoft to supply GPUs for the first Xbox console, thus strengthening its position as a key supplier for the console world, including PlayStation 3.

Another turning point came in 2006 with the introduction of CUDA technology, which allowed GPUs to be used for non-graphics data processing, paving the way for new applications and fields of research, especially for solving mathematical problems using parallel computing.

In 2012, NVIDIA GPUs were used to create AlexNet, a deep neural network. This was the first step towards its current leadership in artificial intelligence and contemporary cloud computing.

Featuring: the OpenAI-Oracle agreement

In September 2025, Wall Street was rocked by incredible news: OpenAI signed a contract with Oracle to purchase $300 billion in computing power over approximately five years.

Beyond economic speculation, given that this is a huge commitment that far exceeds the startup’s revenue at the time of the agreement, this news highlighted two things.

The first is the growth horizon predicted for artificial intelligence services over the next five years, which is expected to continue to be exponential and unrelenting.

The second is the affirmation of Oracle’s OCI service among the giants of cloud computing.

Probably one of the key factors in Oracle’s success in securing this deal is due to the unique features of its offering: payment based on actual resource consumption, seamless integration with Microsoft’s Azure cloud, the possibility of hybrid cloud architecture and complete isolation of its cloud infrastructure, and the choice of hardware used for virtual machine virtualization (Intel, AMD, or Arm).

Whatever Oracle’s trump cards may have been, one thing is certain: today, startups and companies that want to reach an ever-wider audience cannot do without the cloud and its flexibility.

The discordant notes: the AWS US-EAST-1 case

We have extolled its qualities and successes, but the cloud is not infallible. In October 2025, an event reminded the world how fragile the digital infrastructure we rely on every day really is.

On October 20, 2025, the AWS US-EAST-1 region (the busiest and most strategic) suffered a critical failure that caused an outage lasting approximately 15 hours.

The cause was a race condition¹ between two automated systems that managed the DNS configurations of DynamoDB, Amazon’s NoSQL database.

In practice, the two systems attempted to update the same DNS record simultaneously, generating an empty record. This single empty record prevented IP addresses from being resolved, blocking access to DynamoDB and, consequently, to all AWS services that depend on it: EC2, Lambda, EKS, NLB, Cognito, IAM, CloudFormation.

Globally, hundreds of services went offline, including Snapchat, Alexa, Coinbase, Duolingo, and Fortnite, causing disruptions and delays at universities, banks, e-commerce sites, and public administrations. According to CyberCube, estimated losses range from $38 million to $581 million.

Paradoxically, in this case, the cloud showed the same limitations as the central systems it was supposed to solve. In fact, too many services depended on a few cloud regions or a single provider, highlighting the growing need for multi-region and multi-cloud architectures.

The cloud in pop culture: One Piece and Punk Records

Cloud computing, with all its implications, potential, and risks, has also entered pop culture. A prime example comes from One Piece, specifically in the recent Egghead Island story arc, where the brilliant scientist Dr. Vegapunk created a facility called Punk Records.

Thanks to the Nomi Nomi no Mi devil fruit, Vegapunk has a constantly expanding brain. To manage it, he has physically separated his mind from his body, transferring it to an external structure: Punk Records. This “brain cloud” is accessible to Vegapunk and his six satellites, which synchronize with it daily. In this way, he can achieve his goal of creating a shared global archive where anyone can access and contribute new knowledge.

Eiichirō Oda, author of One Piece, with Punk Records, offers a visual and functional metaphor for cloud computing. Specifically:

• Distributed storage: the brain is separated from the body, but accessible remotely.
• Synchronization: satellites update and read data in real time.
• Universal access: Vegapunk dreams of a world where everyone can connect and contribute.

There are also ethical implications. Here is an excerpt from the dialogue between Vegapunk and another character in the manga, Jimbe:

Vegapunk: “If all of humanity were to update their punk records, it would create an ocean of knowledge that would go far beyond my puny brain! One day, humanity will be able to share a single, great brain!”

Jimbe: “But if some ideologies were to get involved, wouldn’t problems arise?”

Vegapunk: “You have a sharp mind, knight of the sea… but… if we had these kinds of scruples, science would never progress!”

As in the real world, doubts also arise in the manga, with Jimbe raising the issue of misinformation and fake news, a clear reference to the contemporary challenges of the modern cloud, to which are added: privacy, security, data control, and the groundness of the information provided by AI.

Conclusions

In this article, we have seen how the cloud has become the invisible infrastructure that powers the digital world. A global orchestra that runs software, distributes intelligence, and connects billions of people. We have touched on some technical concepts and provided real examples and figures. We have highlighted the potential of the cloud and also its critical issues.

And precisely because of these critical issues, just as every symphony requires professionals capable of tuning the instruments, interpreting the score, and ensuring that everything works at its best, in the digital world we need to rely on trained, competent, and thorough professionals.

At noname.solutions, we are part of this symphony.

With our Oracle Cloud Infrastructure (OCI) certifications and expertise in industrial automation, we can help companies design, implement, and optimize scalable, secure, and sustainable cloud solutions.

Because the future isn’t just in the cloud.

It’s in how we play it.

“No one can whistle a symphony. It takes a whole orchestra to play it.”

— Halford E. Luccock, American theologian

👉🏻 Discover our software design and development services.

1. A race condition occurs when the result of an operation depends on the unpredictable sequence in which multiple processes (or threads) access shared resources. ↩︎