[smila-dev] Removed conceptual SCA dependencies

[<Daniel.Stucky@xxxxxxxxxxx>]

Hi all,

I just removed the remaining conceptual SCA dependencies from the
connectivity framework. (I also removed all SCA related comments,
hopefully).

This means, communication between CrawlerController and Crawlers does no
longer rely on Conversations. Instead, I made use of the OSGi
DeclarativeService functionality of so called ComponentFactories for
crawlers. If a crawl is executed, a new instance of the Crawler is
created each time. So it is possible to crawl multiple datasources with
the same Crawler (using multiple instances) in parallel.
This is all fully transparent for the Crawler developer. 

I had to change the usage of the crawler hash for performance counters,
so please check if it still works as desired.

How this ComponentFactory approach is supported by remoting technologies
will be seen, if we have decided for one. For now, this will work fine
for our use cases.

Bye,
Daniel