Networking guides: link to two more solid sources on TCP setting tuning

Author: michaelklishin

docs/networking.md

@@ -593,7 +593,7 @@ using peer certificates is also possible. Please refer to the [TLS/SSL guide](./
 for more information.
 
 
-## Tuning for Throughput {#tuning-for-throughput}
+## Tuning for Throughput and Multi-Gigabit Links {#tuning-for-throughput}
 
 <a id="tuning-for-throughput-intro"></a>
 
@@ -612,7 +612,16 @@ based on message rates and sizes.
 
 For the latter two, see the OS-level tuning section below.
 
-Note that tuning for throughput will involve trade-offs. For example, increasing TCP buffer
+:::tip
+
+[Networking stack tuning](https://fasterdata.es.net/host-tuning/linux/) can be useful for environments with multi-gigabit links and workloads that extensively use [streams](./streams).
+
+However, higher values won't necessarily produce better results, and kernel network
+tuning is [a science, not an art](https://fasterdata.es.net/host-tuning/background/#t2).
+
+:::
+
+TCP setting tuning for throughput will involve trade-offs. For example, increasing TCP buffer
 sizes will increase the amount of RAM used by every connection, which can be a significant
 total server RAM use increase.
 
@@ -884,10 +893,7 @@ Finding an optimal value is usually a matter of trial and error.
 
 ### Nagle's Algorithm ("nodelay") {#tuning-for-large-number-of-connections-nodelay}
 
-Turning off <a
-href="http://en.wikipedia.org/wiki/Nagle's_algorithm">Nagle's
-algorithm</a> is primarily useful for reducing latency but
-can also improve throughput.
+Turning off <a href="http://en.wikipedia.org/wiki/Nagle's_algorithm">Nagle's algorithm</a> is primarily useful for reducing latency but can also improve throughput.
 
 `kernel.inet_default_connect_options` and `kernel.inet_default_listen_options` must
 include `{nodelay, true}` to turn off Nagle's algorithm for inter-node connections.
@@ -1054,6 +1060,15 @@ net.ipv4.tcp_tw_reuse = 1
 
 ## OS Level Tuning {#os-tuning}
 
+:::tip
+
+[Networking stack tuning](https://fasterdata.es.net/host-tuning/linux/) can be useful for environments with multi-gigabit links and workloads that extensively use [streams](./streams).
+
+However, higher values won't necessarily produce better results, and kernel network
+tuning is [a science, not an art](https://fasterdata.es.net/host-tuning/background/#t2).
+
+:::
+
 Operating system settings can affect operation of RabbitMQ.
 Some are directly related to networking (e.g. TCP settings), others
 affect TCP sockets as well as other things (e.g. open file handles limit).

versioned_docs/version-4.0/networking.md

@@ -605,7 +605,16 @@ Tuning for throughput is a common goal. Improvements can be achieved by
 
 For the latter two, see the OS-level tuning section below.
 
-Note that tuning for throughput will involve trade-offs. For example, increasing TCP buffer
+:::tip
+
+[Networking stack tuning](https://fasterdata.es.net/host-tuning/linux/) can be useful for environments with multi-gigabit links and workloads that extensively use [streams](./streams).
+
+However, higher values won't necessarily produce better results, and kernel network
+tuning is [a science, not an art](https://fasterdata.es.net/host-tuning/background/#t2).
+
+:::
+
+TCP setting tuning for throughput will involve trade-offs. For example, increasing TCP buffer
 sizes will increase the amount of RAM used by every connection, which can be a significant
 total server RAM use increase.
 
@@ -1004,6 +1013,15 @@ net.ipv4.tcp_tw_reuse = 1
 
 ## OS Level Tuning {#os-tuning}
 
+:::tip
+
+[Networking stack tuning](https://fasterdata.es.net/host-tuning/linux/) can be useful for environments with multi-gigabit links and workloads that extensively use [streams](./streams).
+
+However, higher values won't necessarily produce better results, and kernel network
+tuning is [a science, not an art](https://fasterdata.es.net/host-tuning/background/#t2).
+
+:::
+
 Operating system settings can affect operation of RabbitMQ.
 Some are directly related to networking (e.g. TCP settings), others
 affect TCP sockets as well as other things (e.g. open file handles limit).

versioned_docs/version-4.1/networking.md

@@ -593,7 +593,7 @@ using peer certificates is also possible. Please refer to the [TLS/SSL guide](./
 for more information.
 
 
-## Tuning for Throughput {#tuning-for-throughput}
+## Tuning for Throughput and Multi-Gigabit Links {#tuning-for-throughput}
 
 <a id="tuning-for-throughput-intro"></a>
 
@@ -612,7 +612,16 @@ based on message rates and sizes.
 
 For the latter two, see the OS-level tuning section below.
 
-Note that tuning for throughput will involve trade-offs. For example, increasing TCP buffer
+:::tip
+
+[Networking stack tuning](https://fasterdata.es.net/host-tuning/linux/) can be useful for environments with multi-gigabit links and workloads that extensively use [streams](./streams).
+
+However, higher values won't necessarily produce better results, and kernel network
+tuning is [a science, not an art](https://fasterdata.es.net/host-tuning/background/#t2).
+
+:::
+
+TCP setting tuning for throughput will involve trade-offs. For example, increasing TCP buffer
 sizes will increase the amount of RAM used by every connection, which can be a significant
 total server RAM use increase.
 
@@ -884,10 +893,7 @@ Finding an optimal value is usually a matter of trial and error.
 
 ### Nagle's Algorithm ("nodelay") {#tuning-for-large-number-of-connections-nodelay}
 
-Turning off <a
-href="http://en.wikipedia.org/wiki/Nagle's_algorithm">Nagle's
-algorithm</a> is primarily useful for reducing latency but
-can also improve throughput.
+Turning off <a href="http://en.wikipedia.org/wiki/Nagle's_algorithm">Nagle's algorithm</a> is primarily useful for reducing latency but can also improve throughput.
 
 `kernel.inet_default_connect_options` and `kernel.inet_default_listen_options` must
 include `{nodelay, true}` to turn off Nagle's algorithm for inter-node connections.
@@ -1054,6 +1060,15 @@ net.ipv4.tcp_tw_reuse = 1
 
 ## OS Level Tuning {#os-tuning}
 
+:::tip
+
+[Networking stack tuning](https://fasterdata.es.net/host-tuning/linux/) can be useful for environments with multi-gigabit links and workloads that extensively use [streams](./streams).
+
+However, higher values won't necessarily produce better results, and kernel network
+tuning is [a science, not an art](https://fasterdata.es.net/host-tuning/background/#t2).
+
+:::
+
 Operating system settings can affect operation of RabbitMQ.
 Some are directly related to networking (e.g. TCP settings), others
 affect TCP sockets as well as other things (e.g. open file handles limit).