What Is Cloud Computing and How Do You Choose a Provider?
Cloud computing is the delivery of computing resources — servers, storage, networking, databases, and software — over the internet on a pay-as-you-go or subscription basis, rather than through hardware you own and run yourself. It fits almost any organization that needs to scale capacity up or down, reach users in many locations, or avoid large upfront infrastructure spending. The harder question is not what cloud computing is but which service and deployment model and which provider fit your workload. This guide walks through the models, the core building blocks, and the selection criteria that actually drive the decision.
The three core service models
The most common way to categorize cloud services is by how much of the stack the provider manages versus how much you manage.
| Model | What the provider manages | What you manage | Typical use |
|---|---|---|---|
| IaaS (Infrastructure as a Service) | Servers, storage, networking, virtualization | OS, middleware, runtime, apps, data | Lift-and-shift migrations, custom stacks, full control needs |
| PaaS (Platform as a Service) | Everything in IaaS plus OS, runtime, and often scaling | Your application code and data | Developers who want to deploy code without managing servers |
| SaaS (Software as a Service) | The entire application | Your data and user configuration | Email, CRM, collaboration, analytics tools |
A concrete example: if you rent a virtual machine and install your own database on it, that's IaaS. If you push code to a managed runtime that handles patching and scaling for you, that's PaaS. If you simply log into a web-based CRM and use it, that's SaaS. The trade-off is consistent — the higher up the stack you go, the less control you have and the less operational work you carry.
Deployment models and when each fits
Service model and deployment model are separate questions. The deployment model describes who the infrastructure serves and where it sits.
- Public cloud — shared infrastructure operated by a provider and offered to many customers. Best when you want elasticity, broad geographic reach, and no capital expenditure.
- Private cloud — infrastructure dedicated to a single organization, either on-premises or hosted. Fits strict control, regulatory, or data-residency requirements.
- Hybrid cloud — a mix of public and private, connected so workloads can move between them. Common when some data must stay on-premises while other workloads benefit from public scale.
- Multi-cloud — using two or more public providers. Chosen to avoid lock-in, meet regional requirements, or use best-of-breed services from different vendors.
These are not mutually exclusive. Many organizations run a hybrid model across two public providers, which is both hybrid and multi-cloud.
Core building blocks to understand
Whatever the model, most cloud platforms expose the same fundamental primitives. Knowing these helps you compare providers on equal terms.
- Compute — virtual machines, containers, and serverless functions. The choice affects how much you manage and how you pay (per-second, per-request, or reserved).
- Storage — object storage for unstructured data, block storage for disks, and file storage for shared filesystems. Each has different durability, latency, and cost characteristics.
- Networking — virtual networks, load balancers, and private connectivity between your environment and the provider.
- CDN and edge delivery — a content delivery network caches content close to users to cut latency. Akamai, for instance, positions itself as both a cloud computing and security company that also operates a large CDN, which matters if your workload is latency-sensitive or globally distributed.
Selection criteria that drive the decision
Provider choice usually comes down to a handful of factors. Weigh them against your specific workload rather than in the abstract.
- Pricing model — understand whether you pay for reserved capacity, on-demand usage, or egress (data leaving the provider). Egress fees are a frequent surprise. Akamai publishes regional pricing pages (North America, Europe, Asia Pacific, São Paulo, and Jakarta), so you can compare rates by region rather than assume a single global price.
- Security — look at the provider's native security services, encryption options, and identity controls. For some buyers, security and cloud are a single evaluation.
- Compliance and data residency — confirm the provider can meet the regulations and geographic restrictions that apply to your data.
- Latency and reach — if users are spread across regions, edge and CDN coverage directly affects performance.
- Lock-in and portability — proprietary services can be convenient but harder to move. Multi-cloud or open standards reduce this risk at the cost of added complexity.
Common migration steps and pitfalls
A typical migration follows a recognizable sequence:
- Assess — inventory workloads and classify them by dependency, sensitivity, and how easily they move.
- Choose a model — decide IaaS, PaaS, or SaaS, and public, private, hybrid, or multi-cloud, per workload rather than for the whole estate at once.
- Pilot — move a low-risk workload first to validate networking, security, and cost assumptions.
- Migrate and optimize — move in waves, then right-size instances and storage to control spend.
Pitfalls to watch for: underestimating egress and data-transfer costs, assuming a lift-and-shift will automatically be cheaper, and skipping the pilot so that integration problems surface only in production. Treat pricing and compliance as things to verify against the provider's own published pages, since rates and terms vary by region and change over time.