What Is Cloud Computing?
Cloud computing means consuming IT resources as on-demand services instead of owning physical infrastructure. This lesson defines the on-demand model, explains virtualization as the technical foundation, and clarifies how AWS delivers services over networks using APIs. Understanding these concepts builds the mental model required to interpret all AWS services correctly.
Cloud computing is the on-demand delivery of IT resources over the internet with pay-as-you-go pricing. The architectural shift is from purchasing servers to consuming services through APIs. In traditional environments, organizations buy hardware, install it in data centers, and manage it for years. In AWS, organizations request resources programmatically and release them when no longer needed.
The core change is operational. Infrastructure becomes software-defined and accessible through API calls.
Note: This model eliminates long procurement cycles and converts infrastructure into a utility.
Defining On-Demand IT Resources
On-demand IT resources are computing components that can be provisioned immediately, used as needed, and released without long-term commitment. In AWS, this includes:
- Virtual servers
- Storage systems
- Databases
- Networking components
Provisioning occurs in minutes or seconds. There is no hardware purchase, shipping delay, or rack installation. This model directly affects cost control and agility. Organizations can align infrastructure usage with real demand.
How AWS Delivers On-Demand Infrastructure
AWS turns traditional IT infrastructure (servers, storage, and databases) into on-demand resources that can be provisioned instantly and released when no longer needed.
For compute, you can launch virtual servers using Amazon EC2 by selecting the instance type, region, and operating system. AWS provisions the instance on shared hardware using virtualization, and you are billed only for the time the instance runs. This provides full OS control and flexibility, while requiring you to manage updates and configuration.
For storage, Amazon S3 provides automatically scalable object storage. You simply create a bucket and upload objects without worrying about disk sizing or hardware management. Storage is billed per gigabyte stored and per request, with extremely high durability and scalability. The trade-off is that object storage behaves differently from traditional file systems and may have higher latency for some workloads.
For databases, Amazon RDS delivers managed relational databases without the overhead of server management. You select the database engine, instance size, and storage configuration, and AWS handles provisioning, backups, patching, and failover. This reduces operational complexity and accelerates deployment, though it limits low-level OS and configuration control compared to self-managed databases.
These services collectively illustrate how AWS abstracts hardware and management tasks, allowing teams to focus on applications and business logic while paying only for what they use.
| Service | Resource Type | What You Control | What AWS Manages | Key Trade-Offs |
|---|---|---|---|---|
| Amazon EC2 | Compute (IaaS) | OS, applications, instance config | Physical hardware | Must handle patching and updates |
| Amazon S3 | Object Storage | Data, access policies | Hardware, scaling, durability | Not a traditional file system; latency may be higher |
| Amazon RDS | Managed Database | Database config, queries | Hardware, backups, patching, failover | Limited OS-level control |
Defining Virtualization
Virtualization allows AWS to divide a physical server into multiple isolated virtual machines. In its data centers, AWS runs a hypervisor on each server to abstract and manage hardware resources such as CPU, memory, storage, and networking.
The hypervisor allocates these resources to create independent virtual machines, such as Amazon EC2 instances. Although these instances run on shared physical infrastructure, each operates as if it has dedicated hardware.
This approach maximizes hardware utilization, supports multi-tenancy, and maintains strong isolation between customers. It also enables rapid, software-driven scaling, making cloud computing both cost-efficient and highly scalable.
The transformation from hardware to virtual resource follows a layered model:
You manage the top two layers while AWS manages the hypervisor and physical hardware.
This layered model defines the Shared Responsibility Model in practice. Understanding this boundary clarifies where your responsibility begins.
Shared Responsibility in AWS
Cloud service models dictate both infrastructure delivery and the allocation of responsibility. Under the Shared Responsibility Model, AWS is responsible for security of the cloud (hardware, networking, virtualization), while customers are responsible for security in the cloud (configurations, identities, data, and applications).
AWS always manages the physical infrastructure and hypervisor layer. Customers never handle hardware failures or data center operations. For example, with Amazon EC2, AWS manages the infrastructure, but you manage the operating system, patches, and security settings. With Amazon RDS, AWS also manages the operating system and database patching, while you manage database access, configuration, and data protection.
When choosing a model, always ask: where does AWS responsibility stop, and where does mine begin?
Network Access Models in AWS
AWS services are delivered over the network through secure, API-driven interactions. You do not physically access AWS infrastructure, instead, you manage resources using the AWS Management Console, AWS CLI, SDKs, or direct API calls. Every action, such as launching an EC2 instance or uploading data to S3, is executed as an API request. This makes AWS infrastructure fully programmable by design.
Most customers access AWS over the public internet using encrypted HTTPS connections secured by TLS. This model provides global accessibility, rapid deployment, and cost-effective connectivity for small to medium workloads. However, because traffic traverses public networks, latency and performance can vary.
For workloads that require more consistent network performance, AWS offers AWS Direct Connect. This service establishes a private, dedicated connection between an organization’s on-premises data center and AWS, bypassing the public internet. Direct Connect delivers more predictable latency and higher bandwidth capacity, making it well suited for performance-sensitive or hybrid environments.
While it requires additional setup and cost, it provides stable, dedicated network performance for enterprise-scale architectures.
APIs as the Control Plane of the Cloud
APIs are the control mechanism of AWS infrastructure. Every AWS service exposes programmatic endpoints. This enables:
- Infrastructure as Code (IaC)
- Automated scaling
- Continuous deployment pipelines
- Event-driven architectures
Cloud environments are programmable platforms designed for automation. Services such as Auto Scaling automatically adjust EC2 instance capacity based on demand, AWS Lambda executes code in response to events like S3 uploads, and AWS CloudFormation provisions entire environments using reusable templates. This automation reduces human error, improves consistency, and enables repeatable, reliable infrastructure deployment at scale.
Control Plane vs. Data Plane
A clear understanding of AWS requires distinguishing between two operational layers. The control plane handles API-driven configuration and management actions, such as launching or modifying resources. The data plane carries the actual workload traffic and processes application data.
For example, launching an EC2 instance is a control plane action, while users accessing a web server running on that instance generate data plane traffic.
This separation enhances reliability, scalability, and operational resilience across the AWS environment.
Architecturally, it ensures management operations do not interfere with workload traffic.
Why This Model Matters for Cloud Practitioners
Cloud computing in AWS is defined by three core principles:
- On-demand resource consumption.
- Virtualized infrastructure abstraction.
- API-driven network delivery.
These principles explain every AWS service you will encounter.
When evaluating any AWS offering, ask:
- Is this an on-demand managed service?
- What layer does AWS manage vs what do I manage?
- How is this service accessed or automated via APIs?
This framework replaces memorization with structured understanding.
