Amazon Web Services, usually abbreviated AWS, is a cloud-computing platform operated by Amazon that provides a large collection of on-demand computing, storage, database, networking, security, analytics, developer, artificial-intelligence and other technology services. AWS began offering infrastructure services in 2006, initially with services including Amazon Simple Storage Service, or S3, and Amazon Elastic Compute Cloud, or EC2. AWS has subsequently developed into a broad cloud ecosystem in which customers can provision individual infrastructure components or assemble complex architectures from managed services. AWS describes itself as a comprehensive cloud platform serving organisations ranging from startups and enterprises to governments and non-profits. The name therefore refers to an extensive portfolio of cloud services rather than a single application, data centre or type of hosting.
AWS provides services across multiple abstraction levels. Amazon EC2 provides configurable compute instances and is broadly associated with IaaS, while Amazon S3 provides object storage and Amazon VPC supplies virtual networking capabilities. Managed database services such as Amazon RDS abstract database administration, while services such as AWS Lambda provide serverless compute. Higher-level services support analytics, machine learning, artificial intelligence, messaging, content delivery, application integration, monitoring, identity and security. AWS's 2024 overview described more than 200 services spanning compute, storage, databases, analytics, networking, developer tools, management, IoT, security and enterprise applications. The platform therefore supports both low-level infrastructure architectures and highly managed application architectures.
AWS is organised into a global infrastructure consisting of geographic regions and separate infrastructure locations within regions. Customers choose regions according to factors such as latency, regulatory requirements, service availability, disaster recovery and cost. Within regions, availability zones provide physically separated infrastructure locations designed to support resilient architectures. A workload can be distributed across multiple availability zones so that a failure affecting one location does not necessarily terminate the entire application.
Cross-region architectures can provide greater geographic separation but require additional design and may introduce network latency, data-transfer costs and data-residency considerations. The global model is one of AWS's central characteristics because it allows applications to be deployed close to users and data while providing mechanisms for geographically distributed architectures.
AWS uses a consumption-oriented commercial model across many services, although pricing mechanisms vary substantially. Resources can be billed according to usage, storage consumed, requests, data transfer, provisioned capacity, reserved commitments or other metrics. This allows organisations to avoid purchasing physical infrastructure for every possible workload level and can support rapid experimentation. AWS's own history describes the service as enabling customers to replace upfront infrastructure expenditure with variable costs that can scale with demand. However, cloud billing can become complex because a single application may consume compute, storage, database, networking, monitoring and other services simultaneously. Effective AWS use therefore requires cost visibility, resource tagging, budgets, rightsizing and architectural awareness of services that generate significant usage or data-transfer charges.
Security is a foundational component of AWS architecture and follows a shared-responsibility model. AWS is responsible for protecting the underlying cloud infrastructure, including physical facilities and foundational systems, while customers remain responsible for security controls appropriate to the services and workloads they operate. AWS provides identity and access management, encryption, key management, logging, network security and other security services that customers can configure. The exact responsibilities differ between services: an EC2 customer manages more of the operating environment than a customer using a fully managed database or SaaS-like service. AWS's security documentation explicitly distinguishes protection of the cloud infrastructure from security implemented within customer environments.
AWS supports infrastructure-as-code and automation through services and tools that allow resources to be defined programmatically. AWS CloudFormation can represent infrastructure in templates, while APIs and command-line tools permit automated provisioning and management. The broader AWS ecosystem supports container orchestration, CI/CD pipelines, monitoring, event-driven architectures and serverless development. These capabilities make AWS suitable for highly automated DevOps workflows in which infrastructure and application deployment are version-controlled and reproducible. Automation can improve consistency and deployment speed but also increases the importance of permission management and testing because a defective template or deployment script can create large numbers of incorrect resources rapidly.
AWS has had substantial influence on the development of modern cloud computing because its early services demonstrated that compute and storage could be consumed programmatically as utility-like services rather than purchased as fixed infrastructure. AWS states that its origins arose from Amazon's own experience with the difficulty and expense of provisioning IT infrastructure, leading to the launch of AWS in 2006. S3 and EC2 became foundational services on which customers could build applications without operating equivalent physical data centres. AWS subsequently expanded into managed databases, serverless computing, analytics, machine learning, security, networking and many other domains. This history illustrates the broader movement from infrastructure ownership towards on-demand cloud services.
AWS should be distinguished from Amazon's consumer retail website and from cloud computing as a generic concept. AWS is Amazon's cloud-services business and platform, while cloud computing describes a broader technological model used by many providers. AWS can provide IaaS, PaaS-like services, serverless services and highly managed application services, so it does not correspond to only one service model. Its significance comes from the breadth of its service catalogue, global infrastructure, programmability and integration among services.
Customers can construct simple virtual servers, highly distributed microservice systems, data-processing pipelines, machine-learning platforms or enterprise architectures from the same overall ecosystem. AWS is consequently best understood as a broad cloud platform and service portfolio rather than a single product or a synonym for “the cloud”.
