Hi Apache Kafka developers! Recently I've been more interested in the difference between APIs and Messaging in a distributed backend? To me, it seems like event-streaming is very fast and is ideal for hooking up all of the non-public facing internal systems of a project, allowing for great decoupling as well. Whereas Apis excel at providing interfaces for external facing systems, but are a bit slower.
However, in my experience, I've worked with tools like graphQL which also provide the great decoupling through the super-graph. It also provides end-to-end type safety, something that seems to be notably lacking from all major messaging systems (as far as I know?). It might be slower, but it's really hard to ignore the thriving ecosystem around api client generation (even for REST/openAPI) and tooling that messaging seems to be lacking. What happened? Is end-to-end type safety and client generation tooling just not a much of a problem in messaging land? I somehow doubt it. Then has APIs won the backend mindshare, when you don't need high throughput, it's easier for most SMBs/SMEs to just use APIs then messaging? That seems... lackluster. I don't see much technical challenges there. I wanted to get some perspective from the developers of Kafka on how they see this ecosystem and trends. No major players in messaging seem to offer end-to-end message type safety, MassTransist, SignalR, Kafka to name a few I've looked for solutions but haven't found anything. And having to have a common code base doesn't work when the distributed services aren't always in the same language. I'd love to hear some feedback on this, Sincerely puzzled and interested, Liam Gryphon
