xref: /freebsd/share/man/man4/mod_cc.4 (revision 2e3f49888ec8851bafb22011533217487764fdb0)
1.\"
2.\" Copyright (c) 2010-2011 The FreeBSD Foundation
3.\"
4.\" This documentation was written at the Centre for Advanced Internet
5.\" Architectures, Swinburne University of Technology, Melbourne, Australia by
6.\" David Hayes and Lawrence Stewart under sponsorship from the FreeBSD
7.\" Foundation.
8.\"
9.\" Redistribution and use in source and binary forms, with or without
10.\" modification, are permitted provided that the following conditions
11.\" are met:
12.\" 1. Redistributions of source code must retain the above copyright
13.\"    notice, this list of conditions and the following disclaimer.
14.\" 2. Redistributions in binary form must reproduce the above copyright
15.\"    notice, this list of conditions and the following disclaimer in the
16.\"    documentation and/or other materials provided with the distribution.
17.\"
18.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
22.\" ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28.\" SUCH DAMAGE.
29.\"
30.Dd September 13, 2022
31.Dt MOD_CC 4
32.Os
33.Sh NAME
34.Nm mod_cc
35.Nd Modular congestion control
36.Sh DESCRIPTION
37The modular congestion control framework allows the TCP implementation to
38dynamically change the congestion control algorithm used by new and existing
39connections.
40Algorithms are identified by a unique
41.Xr ascii 7
42name.
43Algorithm modules can be compiled into the kernel or loaded as kernel modules
44using the
45.Xr kld 4
46facility.
47.Pp
48The default algorithm is CUBIC, and all connections use the default unless
49explicitly overridden using the
50.Dv TCP_CONGESTION
51socket option (see
52.Xr tcp 4
53for details).
54The default can be changed using a
55.Xr sysctl 3
56MIB variable detailed in the
57.Sx MIB Variables
58section below.
59.Pp
60Algorithm specific parameters can be set or queried using the
61.Dv TCP_CCALGOOPT
62socket option (see
63.Xr tcp 4
64for details).
65Callers must pass a pointer to an algorithm specific data, and specify
66its size.
67.Pp
68Unloading a congestion control module will fail if it is used as a
69default by any Vnet.
70When unloading a module, the Vnet default is
71used to switch a connection to an alternate congestion control.
72Note that the new congestion control module may fail to initialize its
73internal memory, if so it will fail the module unload.
74If this occurs often times retrying the unload will succeed since the temporary
75memory shortage as the new CC module malloc's memory, that prevented the
76switch is often transient.
77.Sh MIB Variables
78The framework exposes the following variables in the
79.Va net.inet.tcp.cc
80branch of the
81.Xr sysctl 3
82MIB:
83.Bl -tag -width ".Va hystartplusplus.css_growth_div"
84.It Va available
85Read-only list of currently available congestion control algorithms by name.
86.It Va algorithm
87Returns the current default congestion control algorithm when read, and changes
88the default when set.
89When attempting to change the default algorithm, this variable should be set to
90one of the names listed by the
91.Va net.inet.tcp.cc.available
92MIB variable.
93.It Va abe
94Enable support for RFC 8511,
95which alters the window decrease factor applied to the congestion window in
96response to an ECN congestion signal.
97Refer to individual congestion control man pages to determine if they implement
98support for ABE and for configuration details.
99.It Va abe_frlossreduce
100If non-zero, apply standard beta instead of ABE-beta during ECN-signalled
101congestion recovery episodes if loss also needs to be repaired.
102.It Va hystartplusplus.bblogs
103This boolean controls if black box logging will be done for hystart++ events.
104If set to zero (the default) no logging is performed.
105If set to one then black box logs will be generated on all hystart++ events.
106.It Va hystartplusplus.css_rounds
107This value controls the number of rounds that CSS runs for.
108The default value matches the current internet-draft of 5.
109.It Va hystartplusplus.css_growth_div
110This value controls the divisor applied to slowstart during CSS.
111The default value matches the current internet-draft of 4.
112.It Va hystartplusplus.n_rttsamples
113This value controls how many rtt samples must be collected in each round for
114hystart++ to be active.
115The default value matches the current internet-draft of 8.
116.It Va hystartplusplus.maxrtt_thresh
117This value controls the maximum rtt variance clamp when considering if CSS is needed.
118The default value matches the current internet-draft of 16000 (in microseconds).
119For further explanation please see the internet-draft.
120.It Va hystartplusplus.minrtt_thresh
121This value controls the minimum rtt variance clamp when considering if CSS is needed.
122The default value matches the current internet-draft of 4000 (in microseconds).
123For further explanation please see the internet-draft.
124.El
125.Pp
126Each congestion control module may also expose other MIB variables
127to control their behaviour.
128Note that both NewReno and CUBIC now support Hystart++ based on the version 3 of the internet-draft.
129.Sh Kernel Configuration
130All of the available congestion control modules may also be loaded
131via kernel configuration options.
132A kernel configuration is required to have at least one congestion control
133algorithm built into it via kernel option and a system default specified.
134Compilation of the kernel will fail if these two conditions are not met.
135.Sh Kernel Configuration Options
136The framework exposes the following kernel configuration options.
137.Bl -tag -width ".Va CC_NEWRENO"
138.It Va CC_NEWRENO
139This directive loads the NewReno congestion control algorithm.
140.It Va CC_CUBIC
141This directive loads the CUBIC congestion control algorithm and is included
142in GENERIC by default.
143.It Va CC_VEGAS
144This directive loads the vegas congestion control algorithm, note that
145this algorithm also requires the TCP_HHOOK option as well.
146.It Va CC_CDG
147This directive loads the cdg congestion control algorithm, note that
148this algorithm also requires the TCP_HHOOK option as well.
149.It Va CC_DCTCP
150This directive loads the dctcp congestion control algorithm.
151.It Va CC_HD
152This directive loads the hd congestion control algorithm, note that
153this algorithm also requires the TCP_HHOOK option as well.
154.It Va CC_CHD
155This directive loads the chd congestion control algorithm, note that
156this algorithm also requires the TCP_HHOOK option as well.
157.It Va CC_HTCP
158This directive loads the htcp congestion control algorithm.
159.It Va CC_DEFAULT
160This directive specifies the string that represents the name of the system default algorithm, the GENERIC kernel
161defaults this to CUBIC.
162.El
163.Sh SEE ALSO
164.Xr cc_cdg 4 ,
165.Xr cc_chd 4 ,
166.Xr cc_cubic 4 ,
167.Xr cc_dctcp 4 ,
168.Xr cc_hd 4 ,
169.Xr cc_htcp 4 ,
170.Xr cc_newreno 4 ,
171.Xr cc_vegas 4 ,
172.Xr tcp 4 ,
173.Xr config 5 ,
174.Xr config 8 ,
175.Xr mod_cc 9
176.Sh ACKNOWLEDGEMENTS
177Development and testing of this software were made possible in part by grants
178from the FreeBSD Foundation and Cisco University Research Program Fund at
179Community Foundation Silicon Valley.
180.Sh HISTORY
181The
182.Nm
183modular congestion control framework first appeared in
184.Fx 9.0 .
185.Pp
186The framework was first released in 2007 by James Healy and Lawrence Stewart
187whilst working on the NewTCP research project at Swinburne University of
188Technology's Centre for Advanced Internet Architectures, Melbourne, Australia,
189which was made possible in part by a grant from the Cisco University Research
190Program Fund at Community Foundation Silicon Valley.
191More details are available at:
192.Pp
193http://caia.swin.edu.au/urp/newtcp/
194.Sh AUTHORS
195.An -nosplit
196The
197.Nm
198facility was written by
199.An Lawrence Stewart Aq Mt lstewart@FreeBSD.org ,
200.An James Healy Aq Mt jimmy@deefa.com
201and
202.An David Hayes Aq Mt david.hayes@ieee.org .
203.Pp
204This manual page was written by
205.An David Hayes Aq Mt david.hayes@ieee.org
206and
207.An Lawrence Stewart Aq Mt lstewart@FreeBSD.org .
208