Hands-On Microservices with Kubernetes

Microservices and serverless functions

The same code can often run either as a microservice or as a serverless function. The difference is mostly operational. Let's compare the operational attributes of microservices and serverless functions, as follows:

Microservices	Serverless functions
Always running (it can scale down to at least one). Can expose multiple endpoints (such as HTTP and gRPC). Requires that you implement the request handling and routing yourself. Can listen to events. Service instances can maintain in-memory caches, long-term connections, and sessions. In Kubernetes, microservices are represented directly by the service object.	Runs on demand (theoretically; it can scale down to zero). Exposes a single endpoint ...