From: Andy Furniss <adf.lists@gmail.com>
To: Cake@lists.bufferbloat.net
Subject: cake default target is too low for bbr?
Date: Fri, 28 Apr 2017 20:03:14 +0100 [thread overview]
Message-ID: <7fefc8ad-33e6-cf65-2f90-f71dd2651ee3@gmail.com> (raw)
Motivated by an observation about steam downloads in the thread Getting
Cake to work better with Steam and similar applications I set up a test
- maybe not the best test ever but interesting results.
Put cake on ifb on PC with netem also on egress/ingress adding 4ms each
way (that's how close I am to steam).
Pretend I have 16mbit atm, though here it's shaped on egress and I am
sending 5 connections 10 sec with netperf and bbr enabled.
This doesn't happen with cubic.
OK so ecn may cure - but people may no know how or want that on.
This issue is that there are a lot of drops causing someone with a
highly asymmetric line to possibly use all upload as (s)ack per packet
is prevalent due to excessive drops.
Here's some pastes same test with target varied (via rtt param)
The 5 connections are from different src ips.
qdisc cake 1: dev ifb0 root refcnt 2 bandwidth 16Mbit diffserv3
dual-srchost rtt 100.0ms atm overhead 40 via-ethernet
Sent 17477050 bytes 11597 pkt (dropped 8476, overlimits 31565 requeues 0)
backlog 0b 0p requeues 0
memory used: 144Kb of 4Mb
capacity estimate: 16Mbit
Bulk Best Effort Voice
thresh 1Mbit 16Mbit 4Mbit
target 18.2ms 5.0ms 5.0ms
interval 113.2ms 100.0ms 10.0ms
pk_delay 0us 6.9ms 0us
av_delay 0us 4.6ms 0us
sp_delay 0us 5us 0us
pkts 0 20073 0
bytes 0 30307578 0
way_inds 0 0 0
way_miss 0 10 0
way_cols 0 0 0
drops 0 8476 0
marks 0 0 0
sp_flows 0 5 0
bk_flows 0 1 0
un_flows 0 0 0
max_len 0 1514 0
qdisc cake 1: dev ifb0 root refcnt 2 bandwidth 16Mbit diffserv3
dual-srchost rtt 200.0ms atm overhead 40 via-ethernet
Sent 17459144 bytes 11584 pkt (dropped 3277, overlimits 26349 requeues 0)
backlog 0b 0p requeues 0
memory used: 161280b of 4Mb
capacity estimate: 16Mbit
Bulk Best Effort Voice
thresh 1Mbit 16Mbit 4Mbit
target 18.2ms 10.0ms 10.0ms
interval 208.2ms 200.0ms 20.0ms
pk_delay 0us 12.7ms 0us
av_delay 0us 10.3ms 0us
sp_delay 0us 11us 0us
pkts 0 14861 0
bytes 0 22420426 0
way_inds 0 0 0
way_miss 0 10 0
way_cols 0 0 0
drops 0 3277 0
marks 0 0 0
sp_flows 0 5 0
bk_flows 0 1 0
un_flows 0 0 0
max_len 0 1514 0
qdisc cake 1: dev ifb0 root refcnt 2 bandwidth 16Mbit diffserv3
dual-srchost rtt 300.0ms atm overhead 40 via-ethernet
Sent 17613408 bytes 11680 pkt (dropped 1574, overlimits 24849 requeues 0)
backlog 0b 0p requeues 0
memory used: 171Kb of 4Mb
capacity estimate: 16Mbit
Bulk Best Effort Voice
thresh 1Mbit 16Mbit 4Mbit
target 18.2ms 15.0ms 15.0ms
interval 303.2ms 300.0ms 30.0ms
pk_delay 0us 13.6ms 0us
av_delay 0us 12.1ms 0us
sp_delay 0us 14us 0us
pkts 0 13254 0
bytes 0 19996444 0
way_inds 0 0 0
way_miss 0 10 0
way_cols 0 0 0
drops 0 1574 0
marks 0 0 0
sp_flows 0 5 0
bk_flows 0 1 0
un_flows 0 0 0
max_len 0 1514 0
qdisc cake 1: dev ifb0 root refcnt 2 bandwidth 16Mbit diffserv3
dual-srchost rtt 400.0ms atm overhead 40 via-ethernet
Sent 17674640 bytes 11720 pkt (dropped 920, overlimits 24270 requeues 0)
backlog 0b 0p requeues 0
memory used: 186624b of 4Mb
capacity estimate: 16Mbit
Bulk Best Effort Voice
thresh 1Mbit 16Mbit 4Mbit
target 20.0ms 20.0ms 20.0ms
interval 400.0ms 400.0ms 40.0ms
pk_delay 0us 16.8ms 0us
av_delay 0us 16.2ms 0us
sp_delay 0us 17us 0us
pkts 0 12640 0
bytes 0 19067520 0
way_inds 0 0 0
way_miss 0 10 0
way_cols 0 0 0
drops 0 920 0
marks 0 0 0
sp_flows 0 5 0
bk_flows 0 1 0
un_flows 0 0 0
max_len 0 1514 0
qdisc cake 1: dev ifb0 root refcnt 2 bandwidth 16Mbit diffserv3
dual-srchost rtt 500.0ms atm overhead 40 via-ethernet
Sent 17614080 bytes 11680 pkt (dropped 613, overlimits 23885 requeues 0)
backlog 0b 0p requeues 0
memory used: 186624b of 4Mb
capacity estimate: 16Mbit
Bulk Best Effort Voice
thresh 1Mbit 16Mbit 4Mbit
target 25.0ms 25.0ms 25.0ms
interval 500.0ms 500.0ms 50.0ms
pk_delay 0us 18.6ms 0us
av_delay 0us 17.1ms 0us
sp_delay 0us 19us 0us
pkts 0 12293 0
bytes 0 18542162 0
way_inds 0 0 0
way_miss 0 10 0
way_cols 0 0 0
drops 0 613 0
marks 0 0 0
sp_flows 0 5 0
bk_flows 0 1 0
un_flows 0 0 0
max_len 0 1514 0
next reply other threads:[~2017-04-28 19:03 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-04-28 19:03 Andy Furniss [this message]
2017-04-28 20:45 ` [Cake] " Andy Furniss
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://lists.bufferbloat.net/postorius/lists/cake.lists.bufferbloat.net/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=7fefc8ad-33e6-cf65-2f90-f71dd2651ee3@gmail.com \
--to=adf.lists@gmail.com \
--cc=Cake@lists.bufferbloat.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox