Cloud computing has reduced engineering costs and improved delivery effectiveness drastically over the last decade, but new systems and features can still see lead times of weeks or months even for a prototype release, and operations costs are still often inefficiently managed. Modern approaches in cloud computing, including the new area of serverless architectures, tackle both of these concerns.
Serverless architectures are those that incorporate third-party backend-as-a-service (BaaS) products into the application, or that use functions-as-a-service (FaaS) platforms, like AWS Lambda, to run server-side code in fully managed, event-driven, ephemeral containers. By using these ideas such architectures remove much of the need for the traditional “always on” server system. Depending on the circumstances such architectures can significantly reduce operational cost and increase the speed of experimentation. However the flip side sees extended vendor dependencies and (at present) immature tooling and supporting services.
In this session Mike Roberts gives an introduction to Serverless techniques and proceeds to give cautiously optimistic detail of the benefits and limitations of such an approach, concluding with how he expects Serverless to develop over the coming months and years.