aboutsummaryrefslogtreecommitdiffstats
path: root/docs
Commit message (Collapse)AuthorAgeFilesLines
* dashboard: ignore traffic while suspendedtv5 days1-1/+3
|
* bench tor round tripstv5 days1-1/+1
|
* predecessors go alongtv5 days1-1/+5
|
* covered neighbours have ittv5 days1-0/+8
|
* asked again countertv5 days2-1/+5
|
* neigbours by transporttv5 days2-1/+7
|
* own tree shapetv5 days2-2/+25
|
* drop full mesh stattv5 days2-4/+1
|
* bench mycelium round tripstv5 days1-2/+3
|
* delta from held headstv5 days1-2/+3
|
* offer again ungeneratedtv5 days1-1/+2
|
* stale delta offerstv5 days1-0/+3
|
* delta known to sendertv5 days1-1/+2
|
* delta propagationtv5 days1-1/+2
|
* benchmark bridge fast linktv5 days1-2/+3
|
* benchmark bridge slow linktv5 days1-1/+2
|
* benchmark bridgetv5 days1-1/+4
|
* one delta at a timetv5 days1-1/+5
|
* time blob messagestv5 days3-4/+4
|
* dashboard: hop delaystv5 days1-0/+1
|
* throughput from delivered messagetv5 days1-1/+1
|
* hop delaystv5 days2-0/+33
|
* benchmark members joiningtv5 days1-3/+3
|
* document messages that overtaketv5 days1-4/+20
|
* any message taken as deliveredtv5 days1-7/+9
|
* slow round back offtv5 days2-5/+10
|
* tree messages need no answertv5 days1-3/+10
|
* hello again when heard fromtv5 days1-1/+4
| | | | | | While we wait for a peer's HELLO_REPLY, anything else that comes from it makes our HELLO go again at once, at most once a second, instead of when the backoff says.
* unanswered pull frees the slottv5 days1-3/+4
| | | | | | When the answer to a pull is overdue (the sync reply timeout: 1 s in the simtests, 10 s in production), the pull gives up the slot, and the next waiting neighbour is pulled at once.
* no heal while both move ontv6 days2-2/+2
| | | | | | | The heal clock now also restarts when our heads and the peer's have both changed since the last STATUS that showed them differing. While writes go on, two neighbours's heads differ because of changes still on their way, and that alone no longer leads to a heal.
* pull after gracetv6 days1-1/+3
| | | | | | | | Before pulling, a node waits one measured round trip to the announcer (1 s at most) and pulls only what is still missing. If the round trip isn't known yet, it pulls at once, as now. On the clean benchmark links the round trip is a few milliseconds, so those scenarios shouldn't change.
* announcements answered when lackingtv6 days2-3/+4
| | | | | | | A neighbour now answers an announcement only if it lacks some of the announced heads. Heads that merely differ are no longer enough, since they do while writes go on. The check reads the document, so it runs off the message consumer, as the pull check already does.
* unanswered members not reachedtv6 days2-6/+11
|
* add member away benchmarktv6 days1-1/+2
|
* dashboard: show reach settv6 days1-0/+1
|
* dashboard: show writer treestv6 days2-0/+44
|
* update architecture doctv6 days1-4/+7
|
* writers from authorstv6 days1-5/+7
|
* reach set for the sessiontv6 days2-3/+5
|
* tree changes had alreadytv7 days1-1/+3
|
* announcement round counterstv7 days2-1/+7
|
* docs after treestv7 days1-1/+1
|
* announce what trees misstv7 days1-1/+6
|
* parallel propagation benchmarktv7 days1-1/+3
|
* quicker propagation benchmarktv7 days1-1/+2
|
* order members by distancetv7 days1-4/+6
|
* replace eager links by treestv7 days5-32/+13
|
* add writer treestv7 days1-0/+16
|
* add writer tree benchmarktv7 days1-1/+4
|
* measure 2-hop neighbourhoodtv7 days2-1/+20
|