In order to clarify some of the discussions on this list, I have done some summaries; essentially
a tidied-up write up of issues that were discussed on this list.
Loss Detection:
Currently the CCVal/seqno of packets starting a loss event are out of sync in ccid3_detect_loss,
which hinders fixing other bugs. I made a case analysis for this on
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/detect_loss_algorithm/detect_loss_notes.txt
This contains an entire algorithm, for which a flowchart is on
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/detect_loss_algorithm/detect_loss_flowchart.eps
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/detect_loss_algorithm/detect_loss_flowchart.png
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/detect_loss_algorithm/detect_loss_flowchart.dia
Loss History Handling
A systematic write-up of points raised earlier on this list is on
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/detect_loss_algorithm/summary_of_loss_handling_issues.txt
This includes a list of currently known bugs which need fixing.
Loss History based on Ring Buffer
I have worked out a generic algorithm to achieve fixed-size history handling in a LIFO ring buffer.
The code is deceptively simple, based on a constant HIST_SIZE, and is online on
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/loss_history_ring-buffer/
The detailed description of how this works is on
http://www.erg.abdn.ac.uk/users/gerrit/dccp/docs/loss_history_ring-buffer/circular_fixed-size_loss-history_buffer.txt
I'd welcome any feedback and comments - if there is none, I may end up sending some of
the documents to the list ;-)
-
To unsubscribe from this list: send the line "unsubscribe dccp" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
