Hi everyone!
We're working on an OSGi-based project here and it looks like ACE might be
a good fit.
One of the biggest / hardest requirements is dynamic configuration of
targets. Allow me to elaborate:
- we'll have dozens of OSGi containers running the ACE management agent
- some will pull data by reading local CSV files, others by querying
a database, but all implement the same interface
- each one may need a different set of dependencies
- obviously, the ones that read CSV files will have different
dependencies than the JDBC ones
- different configurations, even if the same "type"
- for example, different JDBC-based clients may run the same code but
might be querying tables with slightly different names
- different DB hostnames, passwords, etc.
- we would like to be able to remotely configure these clients
- e.g., looks like client A needs a different database password
- e.g., looks like client B has the data in CSV files instead of the
database; need to change dependencies
- automatic upgrades
- e.g., if we release a new version of the JDBC code, everyone should
get that
I'm very new to OSGi and ACE (sorry for butchering the vocabulary), but
what attracted us to this project is that it seems to fulfill many of these
requirements already.
What does ACE do, of that list, and what does it not do? Perhaps I'm
barking up the wrong tree entirely and OSGi/ACE is not the right fit.
So, any comments/feedback/suggestions would be greatly appreciated. Thanks
ahead of time!
P.S. We want to implement this as an open source project, either as an ACE
add-on or extension, if it fits into ACE's goals.
--
The Alchemist
