Cloud Computing Concepts
The vocabulary the AWS Certified Cloud Practitioner exam assumes before it asks about any individual service. These are properties of cloud platforms in general; most definitions here follow how Amazon Web Services frames them across its training and exam materials, and each maps onto concrete services such as Amazon EC2 Auto Scaling for elasticity or Elastic Load Balancing (ELB) for fault tolerance.
Key points
- AWS’s official “six advantages of cloud computing” — trade fixed expense for variable expense, benefit from massive economies of scale, stop guessing capacity, increase speed and agility, stop spending money running and maintaining data centers, go global in minutes. This is a specific, named AWS list (see Sources) that elasticity, economies of scale, agility and global reach below map onto loosely; scalability, high availability, fault tolerance and loose coupling are separate exam concepts (closer to the Well-Architected Framework) rather than part of this particular list.
- Service models — three levels of how much of the stack you control:
- IaaS (Infrastructure as a Service): configure and customize the environment from the hardware up. Amazon EC2 is the canonical example.
- PaaS (Platform as a Service): a balance between customization and control — developers focus on the application layer while the platform handles administration, networking and the operating system. AWS Elastic Beanstalk is the canonical example.
- SaaS (Software as a Service): no software installation, the vendor owns everything, the design goal is simplicity and ease of use.
- Deployment models — Public cloud (all resources in a provider’s cloud), Private cloud (dedicated on-premises or hosted infrastructure), Hybrid cloud (the two connected, typically via AWS Direct Connect or AWS Site-to-Site VPN).
- Elasticity — the ability to gain or release compute resources (CPU, RAM, I/O bandwidth, storage) as demand shifts, so peak usage is absorbed without buying equipment for it. Rightsizes resources up and down.
- Scalability — the ability to grow to handle a larger workload. Distinct from elasticity, which is about matching capacity to demand in both directions and doing it automatically.
- Agility — the speed at which resources are provisioned, and therefore the ability to experiment cheaply and abandon failed experiments without stranded capital.
- High availability — an architecture’s ability to withstand failures with minimal downtime, usually by spanning multiple Availability Zones.
- Fault tolerance — the system keeps operating correctly when a component fails, rather than merely recovering quickly.
- Loose coupling — a change or failure in one component must not cascade into others. Achieved with queues and event buses such as Amazon SQS and Amazon EventBridge.
- Economies of scale — aggregate demand across all customers lets the provider drive per-unit prices down continuously, converting fixed costs into lower variable costs.
- Global reach — the ability to deploy infrastructure across multiple geographic Regions connected by a low-latency, high-throughput, redundant network backbone, putting an application near any customer in minutes. Distinct from economies of scale (which is about cost) and elasticity (which is about capacity).
- Rightsizing — continuously matching instance families and sizes to the actual workload, which is the most cost-effective architecture for that workload. See AWS Compute Optimizer.
- Infrastructure as code — automating provisioning by defining infrastructure in templates or code rather than clicking through a console. See AWS CloudFormation and AWS Cloud Development Kit (CDK).
- Authentication vs authorization — authentication defines an identity and verifies it; authorization decides what that identity may do. AWS Identity and Access Management (IAM) handles both.
- Migration strategies — the common exam pair is rehost (lift and shift, unchanged) versus refactor (re-architect, for example converting a monolith into microservices).