Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
update Dandelion documentation #2723
Some specific questions I have about the way this is implemented in Grin:
Yes. One is randomly chosen from the current set of outbound peers at the beginning of each epoch.
Yes. Any stem txs received will be added to the local stempool.
No. I believe we diverge from the spec here. There may be room for improvement to make the behavior match the spec. Did it this way as the impl was significantly simpler if we did not need to treat "local" txs differently. Still not entirely clear what the benefit is of always stemming local txs.
Yes. The following is configurable -
I believe so yes. This implies something unexpected happened in terms of routing and we immediately fluff in this situation. We probably want to keep an eye on this and see how often this occurs in practice.
"Always stem local" behavior implemented here - #2876
Turns out the main thing preventing us from doing this was how hard it makes Dandelion testing if we cannot get "local" txs to participate in local stem aggregation.
So made it a configurable option that defaults to following the Dandelion++ paper.
The following config will set your local node to aggregate all txs it sees and fluff them periodically -
@antiochp thanks for the review, the comments, and the improvement! I've updated the docs to reflect your answers I think and this is now ready to be merged.
Just a quick note regarding this one:
Actually, I don't think the parameters you list here relate to
* generate txhashset archives on 250 block intervals. * moved txhashset_archive_interval to global and added a simple test. * cleaning up the tests and adding license. * increasing cleanup duration to 24 hours to prevent premature deletion of the current txhashset archive * bug fixes and changing request_state to request height using archive_interval. * removing stopstate from chain_test_helper to fix compile issue