Amazon Web Services (AWS) is a subsidiary company of Amazon, offering pay-as-you-go APIs and on-demand cloud computing services. AWS is headquartered in Seattle, Washington, and was founded in 2006. The company offers its services in eighty-four availability zones, in twenty-six geographic regions around the world. AWS's cloud infrastructure is intended to be a scalable and low-cost platform and is used in 190 countries around the world.
AWS was founded in 2006 as a side-business for Amazon and is a cloud infrastructure as a service arm of the company. It was not initially considered largely important, even though in the decade after its launch, AWS grew into the most successful cloud infrastructure company. AWS in 2021 accounted for 32.4 percent of the overall cloud infrastructure market, followed by Microsoft Azure at 20 percent and Google Cloud at 9 percent, and the next largest individual providers. AWS has been compared, in its business model and infrastructure services, with the utility companies that emerged in the early 1900s.
The company offers cloud computing and API products and services for the following:
- Application integration
- Augmented reality
- Virtual reality
- Cost management
- Business applications
- Programming containers
- Customer engagement
- Developer tools
- End-user computing
- Internet of things (IOT)
- Machine learning
- Management and governance
- Media services
- Migration and transfer
- Mobile applications
- Networking and content delivery
- Quantum technologies
- Enterprise software
- Serverless computing
- Data storage
AWS offers its services to customers through a network of server farms located throughout the world, with fees for the services often based on the combination of usage—similar to a "pay-as-you-go" model—with hardware, operating system, software, and networking features chosen by the user. Other factors that can impact the price include the service, availability, redundancy, and security the user wants. AWS also provides some security for users, while it also holds the user responsible for other parts of their security.
The Amazon Web Services communications platform as a Service (CPaaS) is a newer part of the company's platform. It has been developed from strategic partnerships with companies, including Strategic Communications, and through acquisitions, such as the encrypted messaging platform Wickr. These have led to the development of a CPaaS that is directed toward contact centers through cloud deployments. This includes Amazon Connect, which is a customer service cloud contact center offering users tools for agents and managers through a unified interface, and offers features including telephony, high-quality audio, omnichannel outbound campaigns, predictive dialer with answering machine detection, web and mobile chat, omnichannel routing, and a rules engine for contact center automation.
The platform also uses artificial intelligence and machine learning to provide forecasting, capacity planning, and agent scheduling functionality for users, as well as administrative tools and real-time and historical analytics to help optimize the platform. As part of its CPaaS, AWS also offers AWS Communication Developer Services, which includes a set of SDKs and APIs that allow developers to embed customer communications into their applications. This includes application-to-person communications, asynchronous communications, and real-time communications with supported channels that include email, SMS, push notifications, chat, audio, video, and voice.
AWS began offering services and tools to develop generative AI tools and models in 2023. This includes Amazon Bedrock, which offers users a chance to build and scale generative AI applications; AWS Trainium, which allows users to train generative AI models; AWS Inferentia, which allows users to run foundation models (FMs) inferences; Amazon SageMaker, which allows users to build, train, and deploy FMs; and Hugging Face models. Amazon Bedrock does not use generative AI models built by Amazon; rather, it uses third parties to host models on AWS in partnership with various generative AI start-ups. Bedrock offers access to Titan FMs, a family of generative models trained by AWS.
The Amazon Titan FMs offer a range of foundation models for generative AI-based applications. This includes FMs for building text and image-generative AIs. They are pre-trained on large datasets to make them general purpose. Similarly, users of Bedrock can choose from third-party FMs, including the Jurassic-2 family of multilingual LLMs, capable of generating text in Spanish, French, German, Portuguese, Italian, and Dutch; Claude, which can perform a variety of conversational and text processing based on Anthropic's extensive research; and Stability AI's suite of text-to-image foundation models, such as Stable Diffusion, for image generation capabilities.
Bedrock also allows users to develop customized models, using labeled examples in Amazon S3, to help fine-tune the model for a particular task without having to annotate large volumes of data. AWS suggests as few as twenty examples are enough to fine-tune the models.