Cloud computing
Cloud computing is the delivery of computing services — storage, processing, networking, and software — over the internet, on demand, and with elastic scaling. It is often described as a business model or a technology stack, but it is more fundamentally a transformation in how organizations relate to their own infrastructure. Before cloud computing, a company owned its computers; after cloud computing, a company rents capability. The shift is not merely economic; it is ontological. The computer ceases to be a thing and becomes a service.
From Capital to Operating Expense
The traditional model of computing infrastructure required capital expenditure: buy servers, build data centers, hire operators, and amortize costs over years. Cloud computing inverts this model. Through services like AWS, Microsoft Azure, and Google Cloud Platform, organizations pay for what they use, when they use it, and scale up or down without capital commitment. This transformation is the reason startups can launch with almost no infrastructure cost and why enterprises can experiment without multi-year procurement cycles.
But the shift from capital to operating expense is not purely financial. It changes the relationship between engineering and operations. In the old model, infrastructure was a fixed asset that engineering had to work around. In the cloud model, infrastructure is a variable that engineering can manipulate. The Elasticity of cloud resources — the ability to expand and contract capacity in minutes rather than months — means that the infrastructure itself becomes a dynamic variable in the design of systems. This is the real revolution: the infrastructure is no longer a constraint but a control surface.
The Cloud as Distributed System
Cloud computing is, at its core, a distributed system that presents itself as a unified resource. The user sees a single virtual server, a single storage bucket, a single database; behind the interface is a vast, heterogeneous network of physical machines, data centers, and optical fibers, coordinated by software that makes the complexity invisible. This abstraction is powerful but deceptive. The cloud is not a place; it is a set of contractual and technical arrangements that create the illusion of place.
The illusion has limits. Cloud outages — when a single region of a single provider fails — can cascade across the internet because so many services depend on the same underlying infrastructure. The cloud's centralization is its strength and its vulnerability. It offers economies of scale that no individual organization can match, but it also concentrates risk. A single point of failure, when it is as large as AWS us-east-1, is not a point but a continent.
Cloud Native and the New Architecture
The term cloud native refers to systems designed specifically for the cloud model: containerized, dynamically orchestrated, and microservice-based. Cloud native is not merely a deployment strategy; it is an architectural philosophy that treats the cloud's ephemeral nature as a feature rather than a bug. Containers are designed to be created and destroyed quickly. Orchestrators like Kubernetes manage this lifecycle automatically. The system is not designed to be stable; it is designed to be resilient through constant change.
This philosophy has deep connections to general systems theory. A cloud-native system is an open system that maintains its function not by preserving its structure but by continuously replacing its components. It is the computing equivalent of a biological organism that maintains its identity through metabolic turnover. The cloud native architect is not a builder of monuments but a designer of processes — someone who designs for failure, assumes impermanence, and treats every component as disposable.
Cloud computing is the most successful abstraction in the history of technology, and that success is its greatest danger. We have built a system so effective at hiding complexity that we have forgotten the complexity exists. When the cloud fails — and it will — the organizations that depend on it will discover that they have lost not just their infrastructure but their understanding of how infrastructure works. The cloud does not eliminate the need for systems thinking; it merely postpones the consequences of not thinking. The bill comes due in outages, in vendor lock-in, and in the slow atrophy of engineering capability that occurs when a generation grows up never having touched a physical server.