Q&A

What is Cloud Computing?

By AlgoInsight Team2026-06-07awscloud-computingdatabasesnetworking

Overview

Traditionally, running software meant buying servers, hosting them, and maintaining them - expensive, slow to scale, and over- or under-provisioned. Cloud computing replaces that with rented, internet-accessible resources you provision in minutes and pay for by usage. The core promise: turn large up-front capital expense (buying hardware) into flexible operating expense (renting capacity).

graph TD
    A[Cloud Computing] --> B[Service Models]
    A --> C[Deployment Models]
    A --> D[Key Characteristics]
    B --> B1[IaaS: infrastructure]
    B --> B2[PaaS: platform]
    B --> B3[SaaS: software]
    C --> C1[Public] 
    C --> C2[Private]
    C --> C3[Hybrid]
    D --> D1[On-demand & elastic]
    D --> D2[Pay-as-you-go]
    D --> D3[Managed by provider]
    style A fill:#1e1b4b,color:#fff

Service Models (IaaS, PaaS, SaaS)

The classic way to categorize cloud offerings is by how much the provider manages:

The progression IaaS → PaaS → SaaS trades control for convenience: IaaS gives you raw building blocks; SaaS gives you a finished application.

Deployment Models

• Public cloud - shared infrastructure owned by a provider (AWS, Azure, GCP); most cost-effective and scalable.
• Private cloud - dedicated infrastructure for one organization; more control and compliance, higher cost.
• Hybrid cloud - combines public and private, moving workloads between them (e.g. sensitive data on-prem, bursty compute in public cloud).
• Multi-cloud - using multiple providers to avoid lock-in or use each's strengths.

Key Characteristics (NIST definition)

1. On-demand self-service - provision resources without human interaction.
2. Broad network access - available over the internet from anywhere.
3. Resource pooling - providers serve many tenants from shared pools (multi-tenancy).
4. Rapid elasticity - scale out/in quickly to match demand.
5. Measured service - usage is metered; you pay for what you consume.

Benefits

• Cost - no large up-front hardware spend; pay-as-you-go.
• Scalability/elasticity - handle traffic spikes by adding capacity instantly, release it when done.
• Speed/agility - launch resources in minutes.
• Reliability - built-in redundancy, backups, multi-region failover.
• Global reach - deploy close to users worldwide.
• Managed services - offload databases, queues, ML, etc. to the provider.

Trade-offs / Challenges

• Vendor lock-in - heavy use of proprietary services makes switching hard.
• Cost surprises - pay-as-you-go can balloon without monitoring/budgets.
• Security & compliance - a shared responsibility model: the provider secures the cloud, you secure what you put in it.
• Latency / connectivity - dependent on the network; some workloads stay on-prem or at the edge.

Common Cloud Service Categories

Compute (VMs, containers, serverless/Functions), Storage (object, block, file), Databases (relational, NoSQL, managed), Networking (load balancers, CDNs, VPCs), plus AI/ML, analytics, and messaging services.

Why This Matters in Interviews

Cloud questions appear in nearly every modern engineering interview. Know the three service models and the control-vs-convenience trade-off, the deployment models, the NIST characteristics (especially elasticity and pay-as-you-go), and the shared-responsibility security model. Being able to place a concrete product (e.g. "Heroku is PaaS") in the model is a common ask - note this batch's "deploy React to Heroku" question is a PaaS example.

Common Pitfalls

• Confusing IaaS/PaaS/SaaS boundaries.
• Assuming the cloud provider secures your data and config (shared responsibility).
• Ignoring cost monitoring, leading to runaway bills.
• Treating "cloud" as synonymous with one provider rather than a model.

References

• NIST Definition of Cloud Computing (SP 800-145): https://csrc.nist.gov/publications/detail/sp/800-145/final
• AWS "What is Cloud Computing": https://aws.amazon.com/what-is-cloud-computing/

More answers