Dealing with service and choreography dynamics
The number of services to be supported should be unlimited, right?
what is the process to integrate new services into the inventory and use them for a running choreography?
- for example, if we replace a service within the inventory with a new one, should the running choreography be stopped first and re-deploy?
How to deal with temporary service unavailability, such as overloading, breakdown?
- any schemes to include fall-back services with lower QoS?
- timeout setting?