Use Cases for Serverless Apps on the Near and Far Edge

Most of the use cases for near- and far-edge serverless apps boil down to one very simple criterion: If the app needs to be close to the user, consider using the edge. The emphasis here belongs on needs to be closer. Edge computing is always limited more than cloud computing. Sometimes it is cost-effective, but other times (like in Google’s cloud), it may be considerably more expensive to run at the edge. Even more importantly, with only a few exceptions, edge computing tends to not have access to data services like SQL databases, message queues, NoSQL, file storage, and object storage. While Fastly and CloudFlare (both near-edge) have done substantial work here, everyone else lags.

Example tasks for edge computing:

• Rewriting part of an HTTP request in transit (aka “header munging”)
• Managing streaming transformations on media such as images or video
• Ingesting data for training AI or LLMs
• Server Side Rendering (SSR) of web apps
• Running full-fledged stateless microservices
• Computing real time information based on geographic data like weather or location
• Pre-authenticating users

At this point, we have laid a good foundation for understanding serverless apps. In the next chapter, we will take a practical look at building some serverless apps. To do this, we will use Fermyon Spin, an open-source WebAssembly-powered developer tool for serverless apps.