The CDI 2.0 Boot has the ambition to startup CDI
container within application and not opposite as it is in application
servers nowadays.
Do we have any idea how the archive should look like?
It
would be nice if a new layout of directories is developed for
server-less and webcontent-less archives. Similar to WAR or similar to
Spring Bootstrap.
It would be nice if a new
JSR is dedicated and the developer would not necessarily write a code
with SeContainerInitializer used in main() method. This means there
should be some specification declaring how the directory layout is expected and
some Maven plugins with archivers which build this new archive (Boot
Archive or Serverless Archive).
(Similar to WAR + directory for container's jar files and configuration XML)
WDYT?
Next thing is Orchestrator and configurations for JPA/JTA, HTTP Connectors, etc. There should be some place and "standard" schema of the configuration in the archive. Maybe some properties or placeholders in config XML bound with orchestrator or Docker environment. Definitely the static configuration (i.e. IP addresses for e.g. datasources) is not acceptable in Microservices and cannot be embedded in the archive. The next thing of configuration is software multitenancy where static SQL schema cannot be not used from Orchestrator because it changes by tenant connected to REST resource and the tenant is identified by HTTP Header (SQL database schema is the tenant name appeared in the Header as an example).
There should be a big challenge in new EE.
We are going to deal with these problems in my company, so it would be nice if this was already solved in EE.
Cheers
Tibor17
