I'm interesed in design microservices in two different environments, Spring and NodeJs. While in Spring it's easy to find plenty of resources about Netflix Eureka (it's probably the number one with Consul), in NodeJs I found more opportunities to allow microservices to communicate themselves and there isn't a way to go.
These are some methods to design a microservice architecture:
- static ip/port map in code or config file
- through DNS
- Service Discovery (Eureka/Consul)
- P2P (like blockchain?) [also seneca?]
Regarding NodeJs on youtube you can find many videos, one that got my attention is the video of the seneca's father Richard Rodger titled "NodeJs microservices without a registry". The problem with seneca is that to make it work you need a base/main microservice that to me it looks like service discovery since all the other microservices must know its ip and port.
From author website
At the moment, our implementation still depends on “well-known” entry points. You have to run a few base nodes at predetermined locations, so that microservices know where to look to join the network – Peter is fixing that one for us, and soon the network will be completely self-managing.
Maybe after Peter will finish it should look like a P2P microservices architecture, where the knowledge of microservices will be spread with SWIM protocol, it is right?
The other difference is that Seneca use pattern matching to forward the message between microservices, but it always travel through the base/main microservice.
Seneca without service registry, isn't it a service registry itself (know TAGS and where they are instead of IPs)?
Sorry for no code, but I still be in a theorical env at the moment.