Measure how segment size affects upload/download speed. #809
Labels
No Label
0.2.0
0.3.0
0.4.0
0.5.0
0.5.1
0.6.0
0.6.1
0.7.0
0.8.0
0.9.0
1.0.0
1.1.0
1.10.0
1.10.1
1.10.2
1.10a2
1.11.0
1.12.0
1.12.1
1.13.0
1.14.0
1.15.0
1.15.1
1.2.0
1.3.0
1.4.1
1.5.0
1.6.0
1.6.1
1.7.0
1.7.1
1.7β
1.8.0
1.8.1
1.8.2
1.8.3
1.8β
1.9.0
1.9.0-s3branch
1.9.0a1
1.9.0a2
1.9.0b1
1.9.1
1.9.2
1.9.2a1
LeastAuthority.com automation
blocker
cannot reproduce
cloud-branch
code
code-dirnodes
code-encoding
code-frontend
code-frontend-cli
code-frontend-ftp-sftp
code-frontend-magic-folder
code-frontend-web
code-mutable
code-network
code-nodeadmin
code-peerselection
code-storage
contrib
critical
defect
dev-infrastructure
documentation
duplicate
enhancement
fixed
invalid
major
minor
n/a
normal
operational
packaging
somebody else's problem
supercritical
task
trivial
unknown
was already fixed
website
wontfix
worksforme
No Milestone
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Reference: tahoe-lafs/trac-2024-07-25#809
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
#392 has measurements from Brian showing that reducing segment size from 1 MiB to 128 KiB reduced upload/download speed drastically on a high-speed LAN. See if this is still true. Maybe make a little graph of upload/download speed at different segments sizes. There is a script being run by the buildbot to measure upload/download speed, which you could use:
http://allmydata.org/buildbot/builders/speed-colo
If you love this ticket, you might also like #398 (allow users to disable use of helper: direct uploads might be faster) and #320 (add streaming (on-line) upload to HTTP interface).
If you like this ticket, you might also like the "Brian's New Downloader" bundle of tickets: #800 (improve alacrity by downloading only the part of the Merkle Tree that you need), #798 (improve random-access download to retrieve/decrypt less data), #448 (download: speak to as few servers as possible), #287 (download: tolerate lost or missing servers), and #605 (two-hour delay to connect to a grid from Win32, if there are many storage servers unreachable).
Brian's New Downloader is now planned for v1.8.0.
By the way Brian has big plans to make some sort of beautiful graphical visualization of network behavior/performance. I can't wait! For one thing, I intend to work on upload code in v1.8 and such visualization would help tremendously.
Kyle Markley did some good work on this but the results are inconclusive:
More measurements and more analysis of Kyle's measurements (especially his packet traces) is warranted!
See also Nathan Eisenberg's benchmarks of 1.7.1 (standard pipeline and segment size settings) with 1.8.0c1 (standard pipeline and segment size settings):
If #1110 or #1187 (download pipelining) were implemented, then the download performance should in theory be less sensitive to segment size. The segment size could then be reduced to improve memory usage.
Trunk now has the new visualizer in it, so you could run downloads with different segment sizes and then look at the visualizations to see how the segments sizes changed the network behavior.