View source for TCP window scale option

{{Short description|Transmission control protocol configurable option}}
{{Use American English|date=January 2019}}
{{Use mdy dates|date=January 2019}}
The '''TCP window scale option''' is an option to increase the [[receive window]] size allowed in [[Transmission Control Protocol]] above its former maximum value of 65,535 bytes. This TCP option, along with several others, is defined in {{IETF RFC|7323}} which deals with [[long fat network]]s (LFNs).

==TCP windows==
The throughput of a TCP communication is limited by two windows: the [[congestion window]] and the receive window. The congestion window tries not to exceed the capacity of the network ([[congestion control]]); the receive window tries not to exceed the capacity of the receiver to process data ([[Flow control (data)|flow control]]).  The receiver may be overwhelmed by data if for example it is very busy (such as a Web server).  Each TCP segment contains the current value of the receive window.  If, for example, a sender receives an ACK which acknowledges byte 4000 and specifies a receive window of 10000 (bytes), the sender will not send packets after byte 14000, even if the congestion window allows it.

== Theory ==
TCP window scale option is needed for efficient transfer of data when the [[bandwidth-delay product]] (BDP) is greater than 64&nbsp;KB{{binpre|first}}. For instance, if a [[Digital Signal 1|T1]] transmission line of 1.5&nbsp;Mbit/s was used over a satellite link with a 513&nbsp;millisecond [[round-trip time]] (RTT), the bandwidth-delay product is <math>\scriptstyle 1,500,000 \times 0.513 = 769,500</math>&nbsp;bits or about 96,187&nbsp;bytes. Using a maximum buffer size of 64&nbsp;KB{{binpre}} only allows the buffer to be filled to (65,535 / 96,187) = 68% of the theoretical maximum speed of 1.5&nbsp;Mbit/s, or 1.02&nbsp;Mbit/s.

By using the window scale option, the receive window size may be increased up to a maximum value of <math>1,073,725,440\ \scriptstyle \left(= (2^{16}-1)\times(2^{14}) = 65,535 \times 16,384)\right)</math>&nbsp;bytes,  or about 1 GiB.<ref>Borman, D., Braden, B., Jacobson, V., & Scheffenegger, R. (2014). TCP extensions for high performance (No. rfc7323).</ref> This is done by specifying a two byte shift count in the header options field. The true receive window size is left shifted by the value in shift count. A maximum value of 14 may be used for the shift count value. This would allow a single TCP connection to transfer data over the example satellite link at 1.5&nbsp;Mbit/s utilizing all of the available bandwidth.

Essentially, not more than one full transmission window can be transferred within one round-trip time period. The window scale option enables a single TCP connection to fully utilize an LFN with a BDP of up to 1&nbsp;GB, e.g. a 10&nbsp;Gbit/s link with round-trip time of 800&nbsp;ms.

== Possible side effects ==
Because some firewalls do not properly implement TCP Window Scaling, it can cause a user's Internet connection to malfunction intermittently for a few minutes, then appear to start working again for no reason. There is also an issue if a firewall doesn't support the TCP extensions.<ref>{{cite web|url=http://support.microsoft.com/kb/934430 |title=Network connectivity may fail when you try to use Windows Vista behind a firewall device |publisher=Support.microsoft.com |date= |accessdate=2019-07-11}}</ref>

== Configuration of operating systems ==

=== Windows ===
TCP Window Scaling is implemented in [[Microsoft Windows|Windows]] since [[Windows 2000]].<ref>{{cite web|url=http://support.microsoft.com/kb/224829 |title=Description of Windows 2000 and Windows Server 2003 TCP Features |publisher=Support.microsoft.com |date= |accessdate=2019-07-11}}</ref><ref>{{Cite web|url=http://msdn2.microsoft.com/en-us/library/ms819736.aspx|archiveurl=https://web.archive.org/web/20080101144824/http://msdn2.microsoft.com/en-us/library/ms819736.aspx|url-status=dead|title=TCP Receive Window Size and Window Scaling|archivedate=January 1, 2008}}</ref> It is enabled by default in Windows Vista / Server 2008 and newer, but can be turned off manually if required.<ref>{{cite web|url=http://support.microsoft.com/kb/934430|title=Network connectivity fails when you try to use Windows Vista behind a firewall device|date=2009-07-08|publisher=[[Microsoft]]}}</ref>
[[Windows Vista]] and [[Windows 7]] have a fixed default TCP receive buffer of 64&nbsp;kB, scaling up to 16&nbsp;MB through "autotuning", limiting manual TCP tuning over [[long fat network]]s.<ref>{{cite web|url=https://fasterdata.es.net/host-tuning/ms-windows/ |title=MS Windows |publisher=Fasterdata.es.net |date= |accessdate=2019-07-11}}</ref>

=== Linux ===
{{Howto|section|date=February 2016}}
[[Linux kernel]]s (from 2.6.8, August 2004) have enabled TCP Window Scaling by default.  The configuration parameters are found in the [[procfs|/proc filesystem]], see pseudo-file {{mono|1=/proc/sys/net/ipv4/tcp_window_scaling}} and its companions {{mono|1=/proc/sys/net/ipv4/tcp_rmem}} and {{mono|1=/proc/sys/net/ipv4/tcp_wmem}} (more information: {{code|1=man tcp}}, section sysctl).<ref name=linuxsysctl>{{cite web | url = https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt | title = /proc/sys/net/ipv4/* Variables }}</ref>

Scaling can be turned off by issuing the following command. 
{{sxhl|2=console|1=$ sudo sysctl -w "net.ipv4.tcp_window_scaling=0"}} 
To maintain the changes after a restart, include the line {{mono|1="net.ipv4.tcp_window_scaling=0"}} in {{mono|1=/etc/sysctl.conf}} (or {{mono|1=/etc/sysctl.d/99-sysctl.conf}} as of systemd 207).

=== FreeBSD, OpenBSD, NetBSD and Mac OS X ===

Default setting for [[FreeBSD]], [[OpenBSD]], [[NetBSD]] and [[Mac OS X]] is to have window scaling (and other features related to RFC 1323) enabled.<br />
To verify their status, a user can check the value of the "{{mono|1=net.inet.tcp.rfc1323}}" variable via the [[sysctl]] command:
{{sxhl|2=console|1=$ sysctl net.inet.tcp.rfc1323}}

A value of 1 (output "{{samp|1=net.inet.tcp.rfc1323=1}}") means scaling is enabled, 0 means "disabled". If enabled it can be turned off by issuing the command:
{{sxhl|2=console|1=$ sudo sysctl -w net.inet.tcp.rfc1323=0}}

This setting is lost across a system restart.  To ensure that it is set at boot time, add the following line to ''/etc/sysctl.conf'':
<code>net.inet.tcp.rfc1323=0</code>

However, on macOS 10.14 this command provides an error 
 {{samp|sysctl: unknown oid 'net.inet.tcp.rfc1323'}}

== Sources ==
<references />

[[Category:TCP extensions|Window Scale Option]]
[[Category:Network acceleration]]

[[de:TCP Receive Window#TCP Window Scale Option]]