HOSTING / ASSEMBLE FLEXIBLE RESOURCES
Cloud Infrastructure Explained
Cloud infrastructure lets you assemble resources through a console or API. It can scale with demand, but “cloud” alone does not mean automatic scaling, uninterrupted service or a predictable bill.
What You Are Buying
A design may use virtual machines, load balancers, managed databases, object storage and private networks. Each piece has its own failure behavior, quota and price. A single virtual machine in a cloud is still a single machine. Resilience requires a deliberate design, backups and tested recovery; multiple locations add complexity as well as options.
When You Would Use It
- Create and remove environments repeatedly through automation.
- Combine specialized services when that reduces operational work.
- Handle variable demand when the application and spending controls support scaling.
What You Still Manage
The provider runs its infrastructure; your responsibilities depend on the chosen service. You still need account security, access rules, data protection, configuration and a recovery plan. Confirm regions for primary data, backups, logs and support access rather than relying only on a marketing location.
Privacy Considerations
A selected application region may not describe the location of support access, telemetry, logs or every backup. List the separate services involved and ask about their processing locations and subcontractors. Review privileged access and key permissions across the account. A customer-managed encryption key is not automatically a guarantee that the service cannot decrypt data while operating.
Compare provider access, logs, retention and encryption →
Consumption, Allowances & “Unlimited”
Consumption pricing can charge for runtime, provisioned capacity, storage, requests, egress and network services. Reserved commitments trade flexibility for a term obligation. “Pay as you go” does not mean every unused item is free: retained disks, snapshots or allocated services may continue to bill after a VM is stopped.
Read the plan’s resource limits and acceptable-use policy together. Ask what happens at the threshold: an extra charge, reduced speed, rejected requests or suspension. Understand the billing language →
Questions to Ask Before Buying
- Which resources continue charging when my application is idle or stopped?
- Are budget controls alerts, automated actions or enforceable hard caps?
- What quota or capacity could block scaling during a spike?
- What would exporting data and leaving cost?
The takeaway: Elastic capacity and spending control are separate design tasks.
Technical reference: AWS: virtual compute concepts and related resources.
Linked for its explanation of the technology, not as a provider recommendation. Features and contract terms vary.