This post reflects my personal takeaways from Microsoft seminars I’ve attended. It’s designed to help with your revision for MS certifications like AI-900 or AI-102, as I’ve crafted it from my detailed study notes, combined with research I did to better understand cloud computing and AI concepts.
What is cloud computing?
Almost everyone has experienced cloud computing in their daily lives – whether uploading photos and videos to TikTok or Instagram, or tracking health metrics in mobile applications. These services store significant user data, including not only your uploaded content but also information you’ve consented to share such as location and browsing history. This data is transferred securely over the internet and stored in the cloud.
The “cloud” refers to a collection of computing resources provided by cloud service providers such as Microsoft Azure, Amazon AWS, and Google Cloud. It’s much more than just an online USB stick – it’s a full toolbox, not just a drawer. These providers offer comprehensive services that help businesses scale flexibly according to their needs, allowing companies to:
- Create virtual machines without buying physical computers
- Set up entire networks without running cables
- Deploy apps without installing servers
- Run advanced AI systems without specialised hardware
That’s why companies love cloud computing. You can innovate, test, launch – and shut things down – all without purchasing, maintaining, or upgrading physical equipment. The practice of utilising these remote services delivered via the internet is what we call “cloud computing”.
The three core pillars of cloud infrastructure
Cloud infrastructure is established on three essential components:
- Storage: The capability to store, retrieve, and organise data – comparable to memory in a computer system. This includes services like blob storage, file systems, and databases.
- Compute: The processing power that transforms and analyses data – functioning like the brain of the system. This encompasses virtual machines, serverless functions, and container services.
- Network: The connectivity that enables communication between services using the data – similar to a neural network. This includes virtual networks, load balancers, and content delivery systems.
These three fundamental components empower the data stored in the cloud. They enable users to build customised computing environments that can interact with data at higher speeds. Unlike traditional on-premises systems, these cloud resources can be accessed securely from anywhere with an internet connection, at any time.
What types of cloud models exist?
There are three main types — which include one your company, or even your school, might be using.
- Public Cloud
In order to acquire cloud computing power, organisations need to opt for the public cloud – the cloud infrastructure and services made available to tenants who sign up for it. All the services and tools provided by cloud providers like Microsoft Azure, Amazon AWS, and Google Cloud operate within the domain of public cloud.
Imagine wanting to do weight lifting or run an indoor marathon but you can’t afford the equipment. You decide to join a gym instead – everything is already there, so you just sign up and start working out. That’s exactly how public cloud works. If your organisation needs a new IT system, you simply sign up with a cloud provider, and all the infrastructure is ready to be utilised. You just need to login and start building immediately.
By switching to public cloud, companies transform from paying for equipment and hardware sized for imaginary peak capacity to paying only for what they actually use. This shift saves significant resources including initial capital expenditure, planning time, procurement processes, consultant fees, and implementation delays. Capital expenditure becomes operational expense, which is more predictable and economical.
- Private Cloud
Many organisations operate their own private cloud either due to legislative requirements or industry-specific concerns. Banks and healthcare providers, for example, often choose this approach. These organisations are typically financially stable enough to establish and maintain in-house cloud infrastructure, including hardware and software acquisition and maintenance, and they prioritise operating privately due to their possession of sensitive data.
When implementing a private cloud, organisations assume complete responsibility for financing the upfront costs—also known as capital expenditure—as well as setting up data centres, hiring skilled workers, and maintaining the entire system themselves.
This approach, in overall, gives them maximum control over their computing environment and data security, though at significantly higher cost and management complexity compared to public cloud solutions.
- Hybrid Cloud
A hybrid cloud combines both private and public cloud models. For example, an organisation might run their customer-facing application on the public cloud, while keeping their sensitive customer data in a private cloud. This strategic division gives organisations the flexibility and scalability of public cloud services while maintaining the controllability and enhanced security capabilities of a private cloud.
This approach allows businesses to optimise their infrastructure by placing workloads in the most appropriate environment based on performance, security, and cost considerations. Organisations can maintain strict control over mission-critical data while still leveraging public cloud resources for less sensitive operations or to handle demand spikes without over-provisioning their private infrastructure.
Final Thoughts
Cloud computing isn’t just a tech buzzword — it’s powering your apps, your work, and probably your entertainment right now. Whether you’re storing files, launching a website, or training AI — the cloud has your back. So next time someone asks, “What is cloud computing?” — you can tell them the cloud isn’t in the sky, but it’s what’s keeping the world running.