Improve Bandwidth Estimation in Txnsync #3096
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ghost
changed the title
Contributor
|
could you please rebase this PR ? |
tsachiherman
suggested changes
Oct 20, 2021
Contributor
tsachiherman
left a comment
tsachiherman
left a comment
There was a problem hiding this comment.
overall, I think that it looks good. Two small comments -
- The current code is calling
s.node.GetPeerLatencyon every incoming message. That's not horrible wrong, but it would make more sense to cache the latency at thePeerobject, and refresh it only on new round and/or new peer. - You've added quite many debug messages; while there is nothing wrong with that, I don't believe that there is a need to keep these on production code. Specifically for the transaction sync, it has its own logger - and the should be the way to log messages if we find that those are needed. ( I'm not sure if we really would get much value from the log messages added in this PR once we'll be over with these optimizations. ).
Author
|
Yep, the log messages are only in the branch for testing purposes. I will remove them when I'm done testing. |
Codecov Report
@@ Coverage Diff @@
## master #3096 +/- ##
==========================================
- Coverage 43.70% 43.69% -0.02%
==========================================
Files 391 391
Lines 86830 86841 +11
==========================================
- Hits 37946 37942 -4
- Misses 42849 42861 +12
- Partials 6035 6038 +3
Continue to review full report at Codecov.
|
ghost
commented
Oct 20, 2021
ghost
force-pushed
the
nguo/measurelatency
branch
from
3c4033a to
3a1a43e
Compare
ghost
marked this pull request as ready for review
tsachiherman
suggested changes
Oct 22, 2021
Contributor
tsachiherman
left a comment
tsachiherman
left a comment
There was a problem hiding this comment.
Two small nits, but I think that this one is ready to be merged.
Contributor
|
could you please go-lint over the changes ? |
tsachiherman
previously approved these changes
Oct 22, 2021
This reverts commit 3a1a43e.
ghost
force-pushed
the
nguo/measurelatency
branch
from
91f5e80 to
364e328
Compare
This reverts commit 364e328.
onetechnical
pushed a commit
that referenced
this pull request
Oct 25, 2021
## Summary Improve the bandwidth estimation within the transaction sync by having the estimation account for latency, transaction compression time, and time spent waiting in the incoming queue. ## Test Plan Wrote unit tests for correctness, ran network on mainnet model and observed measured bandwidths. Before the bandwidth would converge to the minimum over time as well have erratic inaccuracies. Now the numbers look much more in range, at most a factor of 2 off.
tsachiherman
added a commit
to tsachiherman/go-algorand
that referenced
this pull request
Nov 2, 2021
This reverts commit b93ad64.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Improve the bandwidth estimation within the transaction sync by having the estimation account for latency, transaction compression time, and time spent waiting in the incoming queue.
Test Plan
Wrote unit tests for correctness, ran network on mainnet model and observed measured bandwidths. Before the bandwidth would converge to the minimum over time as well have erratic inaccuracies. Now the numbers look much more in range, at most a factor of 2 off.