You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: _posts/2026/2026-02-16-bonding-iperf3.md
+9-2Lines changed: 9 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -7,7 +7,7 @@ redirect_from: /p/78
7
7
8
8
My workstation in my univerity lab has consistently generated some 600 TiB of annual upload across two PT sites. After an upgrade to the room a few months back, it received the most wanted upgrade for years: A second 1 Gbps line to the campus network. I immediately made the two NICs into one bonded 2 Gbps interface, and it has shined for many times with multiple popular torrents.
9
9
10
-
Inspired by my friend [TheRainstorm](https://blog.yfycloud.site/) who managed to double his upload by load-balancing WireGuard over two lines from the same ISP (China Mobile), I figured I'd leverage this chance and make a better understanding though some more detailed experiments.
10
+
Inspired by my friend [@TheRainstorm](https://blog.yfycloud.site/about/) who managed to double his upload by load-balancing WireGuard over two lines from the same ISP (China Mobile), I figured I'd leverage this chance and make a better understanding though some more detailed experiments.
11
11
12
12
## Setup
13
13
@@ -96,7 +96,14 @@ For 4 parallel streams, the result looks much better:
The gap between RR and TLB no longer exists, and differences in retransmissions and Cwnd size can be attributed entirely to the CC algorithms themselves.
99
-
In particular, due to Cwnd falling under 4 MiB, or maybe just random network fluctuation, retransmission to destination C is observed for the first time.
99
+
In particular, due to Cwnd falling under 4 MiB, or possibly just random network fluctuation, retransmission to destination C is observed for the first time.
100
100
It's also interesting to note that the difference in Cwnd size between BBR and CUBIC increased to 5x, compared to some 3 ~ 3.5x in single-stream case.
101
101
102
102
## Bottom line
103
+
104
+
I did a set of rudimentary experiments in this article, and TCP upload is only a very generic use case.
105
+
For example, @TheRainstorm is running Sunshine streaming inside load-balanced WireGuard tunnel, and turning up forward error correction (FEC) level is one way to offset fluctuation and packet loss.
106
+
I've been running qBittorrent uploads under `balance-tlb` mode for months, and it's been very stable because qBittorrent uploads torrent content over many connections.
107
+
108
+
If anyone wants a more reliable multi-NIC bonding setup, I'd definitely recommend getting a capable switch and doing bonding in `802.3ad` (LACP) mode.
109
+
But for a small homelab, `balance-rr` + BBR would probably suffice for optimizing single-stream transmission speed, at the cost of some overhead on bandwidth.
0 commit comments