Cloud Computing
A model for obtaining configurable computing resources over a network on demand with measured, elastic capacity.
Download Markdown (.md)Soroush’s Product Wiki · https://wiki.soroush.uk/topic/cloud-computing/
01What it means
Cloud computing is a service model in which shared computing capabilities such as servers, storage, networking, platforms, and software are provisioned over a network with limited manual provider interaction. Common service models include infrastructure, platform, and software as a service; deployment may be public, private, community, or hybrid.[1][2]
- Software as a serviceUse a provider-run application
- Platform as a serviceDeploy applications onto managed runtime capabilities
- Infrastructure as a serviceProvision compute, storage, and networking primitives
- Customer responsibilityData, access, configuration, usage, governance
- Provider responsibilityVaries by service and deployment model
Infrastructure, platform, and software service models shift responsibility but never remove customer obligations for identity, data, configuration, governance, resilience, and recovery.
02Why it matters
Product teams can acquire and change capacity faster without owning every physical layer, and can use managed capabilities that would be costly to build. Those benefits exchange capital ownership for recurring cost, provider constraints, shared responsibility, and new skills in governance and operations.
03When to use it
Use cloud services when on-demand provisioning, elasticity, geographic reach, or managed capabilities improve the product and operating model. Dedicated or on-premises infrastructure may fit strict latency, sovereignty, existing investment, disconnected operation, or predictable large-scale economics better.
04How it works
- Select the service and deployment model that matches control, compliance, and team capability needs.
- Provision resources through provider interfaces and preferably repeatable configuration.
- Design identity, networks, data protection, resilience, and observability under shared responsibility.
- Scale or release capacity as demand changes while tracking service quotas and dependencies.
- Meter cost and usage by product, environment, or owner and set governance controls.
- Test portability and recovery assumptions rather than treating provider availability as guaranteed.
05A practical example
A seasonal ticketing product uses managed databases and elastic compute to add capacity before major releases. The team tags resources by product, tests regional recovery, and keeps a cost ceiling because automatic scaling can also amplify inefficient traffic.
06Where it breaks
Cloud services still fail and require skilled operation. Variable bills, data transfer, quotas, jurisdiction, provider-specific services, and migration costs can constrain product choices. Elasticity is not instantaneous or automatically cost-effective.
Common mistakes
- Treating cloud computing as merely running the same servers in someone else's facility.
- Assuming the provider owns application security, backups, and recovery outcomes.
- Using elasticity without budgets, ownership tags, quotas, or cost monitoring.
- Adopting provider-specific services without evaluating exit and continuity needs.
07Sources
- The NIST Definition of Cloud ComputingNational Institute of Standards and Technology · Government standardThe NIST Definition of Cloud Computing
- Cloud Computing Synopsis and RecommendationsNational Institute of Standards and Technology · Government guidanceCloud Computing Synopsis and Recommendations