diff options
Diffstat (limited to 'static/freebsd/man4/tcp_rack.4 3.html')
| -rw-r--r-- | static/freebsd/man4/tcp_rack.4 3.html | 139 |
1 files changed, 139 insertions, 0 deletions
diff --git a/static/freebsd/man4/tcp_rack.4 3.html b/static/freebsd/man4/tcp_rack.4 3.html new file mode 100644 index 00000000..86d73f08 --- /dev/null +++ b/static/freebsd/man4/tcp_rack.4 3.html @@ -0,0 +1,139 @@ +<table class="head"> + <tr> + <td class="head-ltitle">TCP_RACK(4)</td> + <td class="head-vol">Device Drivers Manual</td> + <td class="head-rtitle">TCP_RACK(4)</td> + </tr> +</table> +<div class="manual-text"> +<section class="Sh"> +<h1 class="Sh" id="NAME"><a class="permalink" href="#NAME">NAME</a></h1> +<p class="Pp"><code class="Nm">tcp_rack</code> — <span class="Nd">TCP + RACK-TLP Loss Detection Algorithm for TCP</span></p> +</section> +<section class="Sh"> +<h1 class="Sh" id="SYNOPSIS"><a class="permalink" href="#SYNOPSIS">SYNOPSIS</a></h1> +<p class="Pp">To load the TCP stack as a module at boot time, place the + following line in <a class="Xr">loader.conf(5)</a>:</p> +<div class="Bd Pp Bd-indent Li"> +<pre>tcp_rack_load="YES"</pre> +</div> +<p class="Pp">To enable the TCP stack, place the following line in the + <a class="Xr">sysctl.conf(5)</a>:</p> +<div class="Bd Pp Bd-indent Li"> +<pre>net.inet.tcp.functions_default=rack</pre> +</div> +</section> +<section class="Sh"> +<h1 class="Sh" id="DESCRIPTION"><a class="permalink" href="#DESCRIPTION">DESCRIPTION</a></h1> +<p class="Pp">RACK-TLP uses per-segment transmit timestamps and selective + acknowledgments (SACKs) and has two parts. Recent Acknowledgment (RACK) + starts fast recovery quickly using time-based inferences derived from + acknowledgment (ACK) feedback, and Tail Loss Probe (TLP) leverages RACK and + sends a probe packet to trigger ACK feedback to avoid retransmission timeout + (RTO) events.</p> +<p class="Pp">Compared to the widely used duplicate acknowledgment (DupAck) + threshold approach, RACK-TLP detects losses more efficiently when there are + application-limited flights of data, lost retransmissions, or data packet + reordering events.</p> +<p class="Pp">It is intended to be an alternative to the DupAck threshold + approach.</p> +</section> +<section class="Sh"> +<h1 class="Sh" id="MIB_Variables"><a class="permalink" href="#MIB_Variables">MIB + Variables</a></h1> +<p class="Pp">The algorithm exposes the following scopes in the + <var class="Va">net.inet.tcp.rack</var> branch of the + <a class="Xr">sysctl(3)</a> MIB:</p> +<dl class="Bl-tag"> + <dt id="net.inet.tcp.rack.misc"><var class="Va">net.inet.tcp.rack.misc</var></dt> + <dd>Misc related controls</dd> + <dt id="net.inet.tcp.rack.features"><var class="Va">net.inet.tcp.rack.features</var></dt> + <dd>Feature controls</dd> + <dt id="net.inet.tcp.rack.measure"><var class="Va">net.inet.tcp.rack.measure</var></dt> + <dd>Measure related controls</dd> + <dt id="net.inet.tcp.rack.timers"><var class="Va">net.inet.tcp.rack.timers</var></dt> + <dd>Timer related controls</dd> + <dt id="net.inet.tcp.rack.tlp"><var class="Va">net.inet.tcp.rack.tlp</var></dt> + <dd>TLP and Rack related Controls</dd> + <dt id="net.inet.tcp.rack.timely"><var class="Va">net.inet.tcp.rack.timely</var></dt> + <dd>Rack Timely RTT Controls</dd> + <dt id="net.inet.tcp.rack.hdwr_pacing"><var class="Va">net.inet.tcp.rack.hdwr_pacing</var></dt> + <dd>Pacing related Controls</dd> + <dt id="net.inet.tcp.rack.pacing"><var class="Va">net.inet.tcp.rack.pacing</var></dt> + <dd>Pacing related Controls</dd> + <dt id="net.inet.tcp.rack.tp"><var class="Va">net.inet.tcp.rack.tp</var></dt> + <dd>Rack tracepoint facility</dd> + <dt id="net.inet.tcp.rack.probertt"><var class="Va">net.inet.tcp.rack.probertt</var></dt> + <dd>ProbeRTT related Controls</dd> + <dt id="net.inet.tcp.rack.stats"><var class="Va">net.inet.tcp.rack.stats</var></dt> + <dd>Rack Counters</dd> + <dt id="net.inet.tcp.rack.sack_attack"><var class="Va">net.inet.tcp.rack.sack_attack</var></dt> + <dd>Rack Sack Attack Counters and Controls</dd> +</dl> +<p class="Pp">Besides the variables within the above scopes the following + variables are also exposed in the <var class="Va">net.inet.tcp.rack</var> + branch:</p> +<dl class="Bl-tag"> + <dt id="net.inet.tcp.rack.clear"><var class="Va">net.inet.tcp.rack.clear</var></dt> + <dd>Clear counters</dd> + <dt id="net.inet.tcp.rack.opts"><var class="Va">net.inet.tcp.rack.opts</var></dt> + <dd>RACK Option Stats</dd> + <dt id="net.inet.tcp.rack.outsize"><var class="Va">net.inet.tcp.rack.outsize</var></dt> + <dd>MSS send sizes</dd> + <dt id="net.inet.tcp.rack.req_measure_cnt"><var class="Va">net.inet.tcp.rack.req_measure_cnt</var></dt> + <dd>If doing dynamic pacing, how many measurements must be in before we start + pacing?</dd> + <dt id="net.inet.tcp.rack.use_pacing"><var class="Va">net.inet.tcp.rack.use_pacing</var></dt> + <dd>If set we use pacing, if clear we use only the original burst + mitigation</dd> + <dt id="net.inet.tcp.rack.rate_sample_method"><var class="Va">net.inet.tcp.rack.rate_sample_method</var></dt> + <dd>What method should we use for rate sampling 0=high, 1=low</dd> +</dl> +</section> +<section class="Sh"> +<h1 class="Sh" id="SEE_ALSO"><a class="permalink" href="#SEE_ALSO">SEE + ALSO</a></h1> +<p class="Pp"><a class="Xr">cc_chd(4)</a>, <a class="Xr">cc_cubic(4)</a>, + <a class="Xr">cc_hd(4)</a>, <a class="Xr">cc_htcp(4)</a>, + <a class="Xr">cc_newreno(4)</a>, <a class="Xr">cc_vegas(4)</a>, + <a class="Xr">h_ertt(4)</a>, <a class="Xr">mod_cc(4)</a>, + <a class="Xr">tcp(4)</a>, <a class="Xr">tcp_bbr(4)</a>, + <a class="Xr">mod_cc(9)</a></p> +<p class="Pp"><cite class="Rs"><span class="RsA">Neal Cardwell</span>, + <span class="RsA">Yuchung Cheng</span>, <span class="RsA">Nandita + Dukkipati</span>, and <span class="RsA">Priyaranjan Jha</span>, + <span class="RsT">The RACK-TLP Loss Detection Algorithm for TCP</span>, + <span class="RsD">February 2021</span>, <span class="RsO">RFC + 8985</span>.</cite></p> +<p class="Pp"><cite class="Rs"><span class="RsA">M. Allman</span>, + <span class="RsA">V. Paxson</span>, and <span class="RsA">E. Blanton</span>, + <span class="RsT">TCP Congestion Control</span>, <span class="RsD">September + 2009</span>, <span class="RsO">RFC 5681</span>.</cite></p> +<p class="Pp"><cite class="Rs"><span class="RsA">M. Mathis</span>, + <span class="RsA">Nandita Dukkipati</span>, and <span class="RsA">Yuchung + Cheng</span>, <span class="RsT">Proportional Rate Reduction for TCP</span>, + <span class="RsD">May 2013</span>, <span class="RsO">RFC + 6937</span>.</cite></p> +</section> +<section class="Sh"> +<h1 class="Sh" id="HISTORY"><a class="permalink" href="#HISTORY">HISTORY</a></h1> +<p class="Pp">The <code class="Nm">tcp_rack</code> congestion control module + first appeared in <span class="Ux">FreeBSD 13.0</span>.</p> +</section> +<section class="Sh"> +<h1 class="Sh" id="AUTHORS"><a class="permalink" href="#AUTHORS">AUTHORS</a></h1> +<p class="Pp">The <code class="Nm">tcp_rack</code> congestion control module was + written by <span class="An">Randall Stewart</span> + <<a class="Mt" href="mailto:rrs@FreeBSD.org">rrs@FreeBSD.org</a>> and + sponsored by Netflix, Inc. This manual page was written by + <span class="An">Gordon Bergling</span> + <<a class="Mt" href="mailto:gbe@FreeBSD.org">gbe@FreeBSD.org</a>>.</p> +</section> +</div> +<table class="foot"> + <tr> + <td class="foot-date">March 18, 2024</td> + <td class="foot-os">FreeBSD 15.0</td> + </tr> +</table> |