Start here
Cloud Computing in plain language
A team can request computing capacity through a service instead of procuring a physical machine for each project. Cloud environments may be public, private or hybrid; cloud is not simply another name for the public internet.
Service choice determines the division of work. Renting a virtual machine leaves more operational responsibility with the customer than using a managed application. The useful comparison includes effort, reliability needs, data handling and total cost—not just the advertised compute rate.
How it works
Describe the workload
Record its users, data, availability needs and expected traffic. Decide what the team needs to control.
Choose the service boundary
Compare infrastructure, managed platforms and finished software. Identify responsibilities that remain with the customer.
Operate and review
Monitor usage, access and recovery. Remove unneeded resources and test whether backups can actually restore the required data.
Illustrative example · Fictional data
A temporary testing environment
- Input
- A fictional team needs a test application for a two-week project, using synthetic data.
- Process
- Provision the environment from a repeatable definition, restrict access, set usage alerts and assign an owner to remove it after testing.
- Expected output
- A usable test environment with an explicit end date and a record of the resources created.
- How to check it
- Stopping a virtual machine may leave disks, snapshots and other chargeable resources behind. Verify the entire environment, not just its compute instance.
Where it helps
- Provisioning temporary development or testing capacity.
- Running applications using managed infrastructure or platforms.
- Accessing finished software without operating its underlying servers.
IaaS, PaaS and SaaS
- IaaS
- Rent infrastructure capacity while managing the guest system and application.
- PaaS
- Deploy an application to a managed platform, within its supported runtime and configuration.
- SaaS
- Use a finished application. Customer access, configuration and data responsibilities still remain.
Regions, availability zones and service choices
A region is a geographic deployment area; availability zones provide separate failure domains within a region in services that support them. Distributing a workload across zones can reduce exposure to a single-zone failure, but the application, data replication and failover path must be designed and tested.
A virtual machine, object store and managed database solve different problems. AWS examples include EC2 and S3; Azure offers Virtual Machines and Blob Storage; Google Cloud offers Compute Engine and Cloud Storage. Compare the specific service’s availability, identity controls, data location and costs rather than treating the providers as interchangeable.
Choose a learning and certification path around your role
Begin with identity, networking, compute, storage, billing and shared responsibility on the platform your team uses. A foundation-level certification can organise vocabulary; an administrator, architect or engineer path requires more practical configuration and troubleshooting. An exam pass is not evidence of production readiness on its own.
Use the provider’s current certification catalogue to check the exam name, objectives and retirement notices. Pair study with a small deployment, access review, budget alert and recovery exercise. Do not select a platform solely from an old market-share chart or promised salary premium.
Related reading and learning pathways
Common pitfalls
- Cloud is not automatically cheaper for every workload.
- An available service does not make every application architecture resilient.
- Budget alerts warn about usage; they are not necessarily a hard spending cap.
Put the idea to work
Try it yourself
List every resource a temporary website needs, including storage and logs. Assign an owner and removal check to each item before considering a price estimate.
Common questions
Is cloud computing only public cloud?
No. Public, private and hybrid arrangements exist. The defining idea is the service model and on-demand resource access, not merely the provider name. Many organisations combine on-premises systems with one or more public cloud providers, choosing per workload based on data rules, latency and cost.
Does moving to cloud remove IT operations?
No. Some tasks move to the provider, while application, access, data and cost decisions remain. Teams still manage identity, permissions, application updates, backups, monitoring and spending. The shared-responsibility model in each provider’s documentation sets out who handles what.
Is pay-as-you-go the only pricing model?
No. Services can use usage charges, subscriptions, capacity commitments or combinations. Compare the terms for the specific service. Reserved or committed capacity often lowers unit prices for steady workloads, while pay-as-you-go suits variable demand. Review data-transfer and support charges as well.
Your next step
Continue learning
Connected concepts
Sources and further reading
Examples on this page are teaching exercises using fictional information. They are not client case studies or measured outcomes.