What is Cloud Computing

Cloud computing lets you provision and scale IT infrastructure on demand, without owning hardware. In essence, it is a form of service provisioning. No upfront capital. No lifecycle management. Pay only for what you use, like electricity. You can go from nothing to global infrastructure in minutes.

Origins and History

Year	Milestone
1961	John McCarthy proposes that computing might one day be organized as a public utility
1969	Leonard Kleinrock (ARPANET) predicts the spread of “computer utilities” as networks mature
Early 1990s	Networking industry uses “network cloud” as an abstraction layer for heterogeneous packet-switched networks
Mid-1990s	Public internet services (search, email, social media) validate the core utility computing concept
1999	Salesforce.com pioneers remotely provisioned services directly into the enterprise
2006	Amazon launches AWS and EC2; Google launches Google Apps — the modern commercial era begins

Technology Foundations

Cloud computing is built on several preexisting technologies that serve as both inspiration and building blocks:

Technology	What it is	Cloud contribution
Clustering	Interconnected IT resources acting as a single system; identical hardware, built-in redundancy	Foundation for availability and failover
Grid Computing	Loosely coupled, heterogeneous, geographically dispersed resource pools (a “super virtual computer”)	Inspired resource pooling, load balancing, autonomic configuration
Capacity Planning	Process of determining future IT resource demands to avoid over/under-provisioning	Cloud addresses this via on-demand scaling (lag and match strategies become viable)
Virtualization	Converting physical resources into virtual ones via a hypervisor; severs software dependency on static hardware	Core mechanism for provisioning cloud VMs
Containerization	Highly optimized virtual hosting environments providing only exact required OS resources — no full VM per solution	Lightweight runtime isolation at scale
Serverless Environments	Runtime environments that automatically implement, activate, and scale software packages alongside server components	Eliminates server provisioning entirely for developers

Basic Concepts

IT Resource

An IT resource is any physical or virtual IT-related artifact — a virtual server, a storage volume, a network device, or a software program. When hosted within a cloud, it is a cloud-based IT resource.

On Premises vs. Cloud

An on-premises IT resource is hosted within a conventional enterprise boundary — it cannot simultaneously be a cloud-based resource, but it can interact with cloud resources, be migrated to a cloud, or be deployed redundantly in both environments.

Roles

Every interaction with a cloud involves one or more predefined roles. Organizations and individuals can hold multiple roles simultaneously.

Role	Who it is	Key responsibility
Cloud Provider	Organization that owns and operates the cloud infrastructure	Delivers IT resources per SLA guarantees; handles all infrastructure management
Cloud Consumer	Organization or individual with a formal contract to use cloud resources	Accesses services via a cloud service consumer (app or API client)
Cloud Broker	Third-party intermediary that manages and negotiates cloud services on behalf of consumers	Aggregates services from multiple providers; offers arbitrage and intermediation
Cloud Service Owner	Person or organization that legally owns a cloud service	May be the provider deploying a service or a consumer who builds and publishes their own
Cloud Resource Administrator	Entity responsible for administering a specific cloud-based resource or service	Can belong to the consumer, provider, or a contracted third party
Cloud Auditor	Independent third party that assesses security controls, privacy impacts, and performance	Strengthens trust between consumers and providers through objective evaluation
Cloud Carrier	Network and telecommunication providers responsible for wire-level connectivity	Provides the transport layer connecting consumers to cloud environments

Cloud Service and SLA

A cloud service is any IT resource made remotely accessible via a cloud — from a simple web application to an administrative API. The usage conditions, behaviors, limitations, and Quality of Service (QoS) features of a cloud service are defined in a Service-Level Agreement (SLA): the contractual guarantee of uptime, reliability, and performance.

Boundaries

Cloud environments operate within overlapping perimeters that determine ownership and trust:

Organizational Boundary — the physical perimeter surrounding the IT assets owned and governed by a specific organization. Both cloud consumers and cloud providers have their own distinct organizational boundaries
Trust Boundary — a logical perimeter that extends beyond physical limits to represent the extent to which an organization trusts external IT resources. When a consumer moves workloads to the cloud, they must extend their trust boundary to include the provider’s environment
Logical Network Perimeter — a security boundary established within a cloud environment that isolates a consumer’s network from other tenants and from the broader internet; implemented as a cloud infrastructure mechanism

Business Drivers

Two fundamental drivers motivated the creation and adoption of cloud computing:

Driver	Problem it solves
Cost Reduction	Traditional IT requires massive upfront capital investment; operational overhead (staff, patches, power, cooling, security) often exceeds hardware costs. Cloud replaces capital expenditure with measured operational usage
Business Agility	Organizations with rigid on-premises infrastructure cannot respond to usage fluctuations or launch new initiatives quickly. Cloud enables instant resource scaling and faster solution deployment

Goals and Benefits

Increased Responsiveness

On-demand scalability, high data availability, and reduced infrastructure maintenance enable employees to work remotely and organizations to release solutions faster
Eliminates the internal administrative burden of maintaining platforms — the cloud provider handles it
Significantly reduces time-to-market for new business initiatives

Reduced Investments and Proportional Costs

Public cloud providers purchase IT resources in bulk and pass economies of scale to consumers
Replace anticipated capital expenditure with measured operational expenditure that scales with business performance
Fine-grained resource control — add or remove compute, storage, and bandwidth in small increments
The “as a service” advantage: technical complexity is abstracted into ready-to-use solutions, reducing the IT expertise needed to operate them

Increased Scalability

Clouds instantly allocate pooled resources either on-demand or through consumer-defined configuration
Automated elasticity handles unpredictable usage peaks without manual intervention — tied directly to proportional costs
Enables lag and match capacity planning strategies instead of costly lead-based over-provisioning

Increased Availability and Reliability

Cloud environments are designed from the ground up to minimize outages through modular, redundant architecture
Extensive failover support enables resources to avoid and recover from runtime failures faster than traditional infrastructure
Ultimately governed by the provider’s SLA — availability guarantees are contractual obligations, not engineering aspirations

Scaling

Horizontal Scaling (Scale Out / In)

Allocate (scale out) or release (scale in) additional IT resources of the same type
Common in cloud environments — uses commodity hardware, no single point of failure
Not limited by maximum hardware capacity; requires automated scaling mechanisms
Lower cost per unit; well-suited to stateless workloads

Vertical Scaling (Scale Up / Down)

Replace an existing resource with one of higher (scale up) or lower (scale down) capacity
Less common in cloud — typically requires downtime during replacement
Limited by maximum hardware capacity; more expensive per unit; suited to stateful or single-instance workloads

Load Balancing

When using horizontal scaling, a load balancer distributes incoming traffic across healthy instances using algorithms including:

Round Robin — cycles sequentially through available instances
Least Connections — routes to the instance with the fewest active connections
Resource Utilization — routes to the instance with the lowest CPU or memory utilization

Autoscaling

The automated process of adding or removing instances in response to usage demand or resource utilization. Configure thresholds — connection counts, CPU utilization percentages — that trigger the cloud provider to provision or terminate instances automatically.

Serverless

Serverless eliminates the need to provision or manage servers. The cloud provider automatically scales execution environments per invocation.

Original model (e.g., AWS Lambda): Deploy code as functions; the provider handles all infrastructure; billing is strictly per execution
Modern usage: Cloud providers increasingly label services as “serverless” when they auto-scale, even if billing is still infrastructure-based. Evaluate what “serverless” means in each specific service before assuming zero-idle costs

Risks and Challenges

Risk	What it means
Overlapping Trust Boundaries	Moving data to the cloud expands your trust boundary to include the provider’s environment. In public clouds, resources are shared across multiple consumers — increasing attack surface
Shared Security Responsibility	Security is divided between provider and consumer. The provider typically secures the infrastructure; the consumer is responsible for data, identity, and application-layer security
Increased Cyber Threat Exposure	Internet-accessible resources face a broader threat surface than on-premises systems. Ultimate responsibility for risk management lies with the consumer
Reduced Operational Governance Control	Consumers have lower governance over cloud resources — reliance on provider SLAs, network latency to remote data centers, and external operational decisions introduce risk
Limited Portability	Proprietary APIs and custom-built integrations create lock-in. Moving workloads between cloud providers is technically complex and often costly
Multiregional Compliance	Data may reside in jurisdictions with conflicting privacy laws. Regulatory frameworks (GDPR, HIPAA, etc.) hold the consumer ultimately responsible regardless of where the provider stores data
Cost Overruns	Migrating without a proper business case leads to poor governance and unexpected capital requirements. On-demand provisioning makes it easy to accumulate spend without visibility

Delivery and Deployment Models

This page covers what cloud computing is. For the detailed breakdown of service and deployment models:

Cloud Delivery Models — IaaS, PaaS, SaaS — comparison, combining models, and provider vs consumer perspectives
Cloud Deployment Models — public, private, hybrid, and multicloud trade-offs