Do you have more information about the bottleneck you see?
We have an update for the Event Admin specification (see RFC 157 in [1]) to address some performance issues. But it is not completed yet.
Send an event is a synchronous send, so the calling thread is blocked until all listeners are notified. Most implementations use a caller stream. The message does not block the caller. It pauses for another thread to deliver the event. The current event manager specification requires ordering for asynchronous events, so this can lead to delays if an order is not required. RFC 157 provides the option to not require this order.
I have no data on whether one implementation was better or worse than another.
[1] http://www.osgi.org/Download/File?url=/download/osgi-4.3-early-draft2.pdf
source share