Design distributed software systems

A short brain dump after having an intense discussion. First, what is a distributed system? We agreed, that two processes which do not share a „self“ controlled scheduler e.g., the Linux „systemd“, can be seen as a distributed system. If so and according to the used execution manager behaviour, each middleware/ runtime can be seen as a distributed system.

Under these constraints the todays microservice approaches can be 1:1 applied to service oriented middleware… and due to that for Software Defined Vehicle. We have to split service into fine granular functions with distributed transactions e.g. the shared state approaches, side cars (uP to UP, uC to uP, …), etc.