This pull request includes a newly developed, more flexible interface to the online analysis. It achieves true zero copy data distribution to analysis processes by implementing a publish-subscribe mechanism on top of shared memory ipc and reference counting. The online analysis worker processes can be started, stopped and restarted independently, and they can specify different queueing modes and subsampling parameters.
This pull request includes a newly developed, more flexible interface to the online analysis. It achieves true zero copy data distribution to analysis processes by implementing a publish-subscribe mechanism on top of shared memory ipc and reference counting. The online analysis worker processes can be started, stopped and restarted independently, and they can specify different queueing modes and subsampling parameters.