Compare all serverless web and mobile architecture patterns with standard PoVs in design, impact, motivation, and best practices. These patterns can coexist.

This article is part of a series of articles that examine serverless architecture from different perspectives and provide practical guidelines to help adopt serverless architecture, solve practical challenges in serverless architecture, and discuss how serverless supports reactive event driven architecture. These articles do not mention the serverless services of cloud providers, but only refer to the services in the examples (AWS is a common reference).

As discussed in the article "Serverless Reference Architecture", serverless has its own footprint in different workload modes. The Web is the most common and ubiquitous type of application that allows content and business travel to be delivered over the Internet. This article introduces five modes of serverless Web architecture in detail, and describes the motivation and impact of using this mode in detail. These designs are presented in a cloud/vendor neutral manner with complementary AWS implementations. Considering the service scope of all major cloud providers, the same model can be implemented in all services. Finally, the details include some potential pitfalls and best practices.

Some cross concerns in these modes are:

Idempotency of serverless functions Potential problems of stateless workloads and concerns about cold start Use message boundaries such as dead letter queues

Depending on the context and requirements, these patterns can coexist in the same application.