View source for Hardware: ZyXEL PLA-4201 Twin Pack - throughput test

It is not so easy to add wires/connection once flat or house is build. Even when people think about devices placement they can end up with missing Ethernet connection for some device. WiFi connection is not usable for all kind of network communication due to nature of it. WiFi have limited throughput, jitter and other kind of problems caused by "wireless" connection. So is powerline connection solution to this problem ? At beginning of this document i can say: "I don't know", but i try to find answer.

I bought [http://www.zyxel.com/uk/en/products_services/pla4201.shtml?t=p ZyXEL PLA-4201] recently and I use this device during tests to find answers. I know how this device work by theory but I don't know what is in real word. In theory it should modulate all data to mains with hope that second device "catch" those data without too much distortion but there are several problems:
* switching power supplies and power regulators (light dimmer, vacuum cleaner "speed" settings, etc.) are creating serious noise covering wide range of frequencies
* power cables are not optimized to transfer "data". Most of the filters kill anything else that 50/60 Hz.
* if sockets are not on same power line then communication (as expected) is not possible
* communication can't pass transformers, surge protectors and other similar devices

Power cables are definitely not "polite" place for Ethernet data, but why don't test it ? In office I heard opinions like: "it will work only on 10 meter cable or shorter", "noise kill all transferred data", etc. Will it be true ?

I prepared several tests. Here is description of tests:
== Methodology of synthetic tests ==
First of all I waited for everything to settle down and then at each test point I executed following tests. I run first test once more if it result in high response rate. It looks like that ZyXEL powerline adapters needs some traffic to "sync" and configure connection.

=== Test #1 ===
<pre>ping -c 10 server</pre>
Simple ping test to see reaction of connection on as little traffic as possible (no other communication was flowing thought extenders during this and also other tests)

=== Test #2 ===
<pre>ping -c 100000 -f server</pre>
Ping food test with default ping size. This test is generating relatively small packets at speed of source network card. Purpose of this test is to see packet lost and response time if queuing occur.

=== Test #3 ===
<pre>ping -c 100000 -f -s 1400 server</pre>
Same as Test #3 but with size close to MTU. I selected smaller size as devices from some vendors have great throughput at size of MTU but can't maintain it when packed size is smaller.

=== Test #4 ===
<pre>netperf -l 60 -H host</pre>
"One minute" classic netperf test to see maximu throughput.

=== Test #5 ===
<pre> netperf -t omni -j -l 60 -H host -- -d rr</pre>
Omni netperf test mainly to see number of transactions but other results of this tests are also interesting.

=== Test #6 ===
<pre> netperf -t omni -j -l 60 -H host  -- -d rr -O \
"MIN_LATENCY,MAX_LATENCY,P50_LATENCY,P90_LATENCY,P99_LATENCY,MEAN_LATENCY,STDDEV_LATENCY"</pre>
Same omni test as in #5 but configured to show latency. I expect that test #2 and #3 show min/avg/max but it is also interesting to see this via percentile calculation. (e.g. several delayed packets can affect average calculation but they will be omitted in percentile output)

== Direct connection test ==
First of all I need reference data. I connected two included 1 meter long Ethernet cables together and connected them directly to server's wire. Other end i connected to my laptop.

[[File:ZyXEL PLA-4201 - direct connection.jpg]]

Computers negotiated 1000baseTx-FD with flow-control but later i discovered that ZyXEL PLA-4201 have only 100baseTx-FD (flow-control)  connection. Package of this product is little bit misleading. It claims "500Mbps" throughput and support for 10/100/1000baseTx but Ethernet port on device is 10/100baseTx only. 500Mbps technology mean that you have more "power" to survive interference and one "powerline" segment can handle higher throughput resulting in seamless communication between several devices at same time. Tested powerline device provide hardware flow control at Ethernet ports. This is very good benefit as computer know possible transmission speed depending of powerline status.

I tested using following devices:

Server:
* Asrock ION-330HT
** lan: NVIDIA Corporation MCP79 Ethernet

Client:
* HP ProBook 6555b
** lan: Marvell Technology Group Ltd. Yukon Optima 88E8059 [PCIe Gigabit Ethernet Controller with AVB]
** wlan: Broadcom Corporation BCM43224 802.11a/b/g/n (rev 01)

Wifi access point:
* Asus RT-N10U
** Firmware: DD-WRT v24-sp2 (02/11/13) big
** Network: N-Only (40MHz wide channel, but due to limitation of connected devices only about 65 Mbps raw speed was used)
** Encryption: WPA2/AES with recommended "N Only" configuration

=== 100baseTx-FD ===
Direct connection using CAT5 cable can provide excellent response time. By help of this we can see almost 2500 Transactions per seconds. Throughput is also very good. Daemons on both sides are running in user-space so this measurement include delay caused by driver, kernel and process scheduling.

Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 8998ms
rtt min/avg/max/mdev = 0.304/0.365/0.442/0.038 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 4248ms
rtt min/avg/max/mdev = 0.206/0.300/0.555/0.032 ms, ipg/ewma 0.424/0.303 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 7185ms
rtt min/avg/max/mdev = 0.480/0.579/0.849/0.042 ms, ipg/ewma 0.718/0.562 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.27      94.15   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   2445.96    Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
209          3627         410          429          486          405.87       41.78        
</pre>

=== WiFi ===
As i mentioned earlier WiFi have limited thought clearly show by Test 4 and 5. Response time is good while there is limited amount of traffic. By saturating connection response time jump to so high numbers. Response time at around 200 ms can cause noticeable delay in applications. For real time applications this can be serious problem.

Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9012ms
rtt min/avg/max/mdev = 1.059/1.140/1.221/0.060 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 11760ms
rtt min/avg/max/mdev = 0.782/1.054/13.972/0.360 ms, pipe 2, ipg/ewma 1.176/1.062 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 9998 received, 0% packet loss, time 22377ms
rtt min/avg/max/mdev = 1.575/2.506/218.434/8.988 ms, pipe 5, ipg/ewma 2.237/1.732 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.06      39.42   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   800.10     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
805          360236       1082         1185         2896         1176.57      3287.30      
</pre>

== Power line setup ==
At first i connected devices as close as possible to have reference data. Only UPS was connected to used wiring as it is powering core components of my flat and it was not possible to disable it during tests.

[[File:ZyXEL PLA-4201 - side by side.jpg]]

Then i removed one device and connected test wiring.

[[File:ZyXEL PLA-4201 - point 0 (uplink).jpg]]

At end of 7 meter long cable i connected hair dryier.

[[File:ZyXEL PLA-4201 - point 1.jpg]]

At end of 5 meter long cable i connected saving light bulb.

[[File:ZyXEL PLA-4201 - point 2.jpg]]

At end of 1.5 meter long cable i connected vacuum cleaner.

[[File:ZyXEL PLA-4201 - point 3.jpg]]

At end of another 5 meter long cable i connected saving light bulb. This line contain only two wires. Normally it should have tree wires but powerline devices from ZyXEL use only two wires so it should not be problem.

[[File:ZyXEL PLA-4201 - point 4.jpg]]

As last connection i used 5 meter long cable designed for reading lamps.

[[File:ZyXEL PLA-4201 - point 5.jpg]]

== Power line tests ==
=== "Side by Side" ===
During this test i disconnected all selected "noise provides" and tested devices while there was no other power consumption on testing wiring. I was little bit surprised by ~3.5 ms delay during this test. I repeat first tests several times and also checked configuration of connected devices. I expect better response time but it looks like it is caused by technology itself (conversion from Ethernet to powerline and then back). While i was repeating tests i noticed also stability of this response. Throughput test show excellent result (almost as with direct Ethernet connection), but number of transactions are less due to delay on line.

Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.523/3.735/5.223/0.501 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 36176ms
rtt min/avg/max/mdev = 3.242/3.500/7.530/0.429 ms, ipg/ewma 3.617/3.608 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 132008ms
rtt min/avg/max/mdev = 6.744/22.551/38.489/1.688 ms, pipe 4, ipg/ewma 13.202/22.842 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.27      93.99   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   618.73     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1309         13092        1557         1670         4234         1625.58      398.72       
</pre>

=== "Alone on line" ===
Values measured at end of ~24 meter long cable was almost similar to values measured on few centimeter long wire. Differences are under expected errors during measurement. It looks like distance is not so big problem while there is no noise on line. Throughput at speed of 100baseTx-FD connection and delay only ~3.5 ms is good result. Unfortunately currently most of the customers devices use switching power supply. Even if this is benefit for application itself (smaller size, less heat dispatched through power supply, stable output, etc.) they are source of serious noise. Results from this test can be marked as "lab" results. (Note: Saving Light Bulbs are also source of serious noise, especially if they are "cheap china" model)

==== Test point #1 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.444/3.652/5.306/0.554 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 35869ms
rtt min/avg/max/mdev = 3.222/3.481/8.522/0.416 ms, ipg/ewma 3.587/3.558 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 137557ms
rtt min/avg/max/mdev = 12.323/22.494/38.041/1.444 ms, pipe 4, ipg/ewma 13.757/22.311 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.30      93.95   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   618.60     Trans/s  
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1304         8705         1559         1676         3800         1610.93      315.40       
</pre>

==== Test point #2 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9014ms
rtt min/avg/max/mdev = 3.353/3.450/3.494/0.085 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 35962ms
rtt min/avg/max/mdev = 3.265/3.493/11.572/0.445 ms, ipg/ewma 3.596/3.694 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 136178ms
rtt min/avg/max/mdev = 6.542/22.470/37.448/1.574 ms, pipe 4, ipg/ewma 13.619/22.390 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.33      94.00   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   618.98     Trans/s   
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1301         6588         1559         1678         3725         1613.09      313.37       
</pre>

==== Test point #3 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.449/3.477/3.532/0.084 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 35758ms
rtt min/avg/max/mdev = 3.252/3.470/9.408/0.404 ms, ipg/ewma 3.576/3.381 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 131589ms
rtt min/avg/max/mdev = 6.439/22.533/35.957/1.583 ms, pipe 4, ipg/ewma 13.160/22.297 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.33      93.99   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   618.56     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1306         7065         1560         1677         4315         1616.12      337.55       
</pre>

==== Test point #4 ====
Test 1:
<pre>
--- 10.54.1.7 ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9014ms
rtt min/avg/max/mdev = 3.376/3.470/3.552/0.074 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 36449ms
rtt min/avg/max/mdev = 2.897/3.540/15.270/0.539 ms, pipe 2, ipg/ewma 3.645/3.478 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 128001ms
rtt min/avg/max/mdev = 12.291/22.570/38.633/1.883 ms, pipe 4, ipg/ewma 12.801/21.577 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.28      94.05   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   615.94     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1304         5839         1560         1682         3513         1618.76      314.13       
</pre>

==== Test point #5 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9014ms
rtt min/avg/max/mdev = 3.402/3.513/4.071/0.193 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 36245ms
rtt min/avg/max/mdev = 2.724/3.520/20.094/0.523 ms, ipg/ewma 3.624/3.408 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 125892ms
rtt min/avg/max/mdev = 6.197/22.571/38.242/1.885 ms, pipe 4, ipg/ewma 12.590/22.749 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.25      94.00   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   617.92     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1305         8036         1560         1679         3663         1615.62      320.26       
</pre>

=== "Noise on line" ===
I was not expecting it. Tested devices are robust to survive serious noise and also length of cable. I expected degradation of speed but also problems with response time. Even when there was only half of the throughput available response time was almost unchanged. Also there was no packed lost at all. Packet lost can slow down applications so it is good to know that packed lost don't occur even under difficult conditions.

==== Test point #1 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.461/3.910/4.546/0.510 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 37725ms
rtt min/avg/max/mdev = 2.547/3.670/9.925/0.619 ms, ipg/ewma 3.772/3.534 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 132143ms
rtt min/avg/max/mdev = 5.947/22.587/43.401/2.914 ms, pipe 4, ipg/ewma 13.215/23.811 ms
<pre>

Test 4:
<pre>
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.34      93.50   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   594.37     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1310         9838         1570         2201         3004         1688.51      362.65       
</pre>

==== Test point #2 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.409/3.732/4.645/0.450 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 38524ms
rtt min/avg/max/mdev = 3.269/3.749/10.248/0.667 ms, ipg/ewma 3.852/3.747 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 137121ms
rtt min/avg/max/mdev = 6.111/22.856/42.547/3.892 ms, pipe 4, ipg/ewma 13.713/22.175 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.24      91.68   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   581.57     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1326         7898         1581         2210         3714         1705.86      386.82       
</pre>

==== Test point #3 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.453/3.748/4.536/0.415 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 39535ms
rtt min/avg/max/mdev = 2.414/3.848/9.581/0.785 ms, ipg/ewma 3.953/4.099 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 136755ms
rtt min/avg/max/mdev = 3.359/22.937/38.814/4.030 ms, pipe 4, ipg/ewma 13.676/21.981 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.36      77.04   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   564.57     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1381         9904         1600         2284         4480         1773.36      511.98       
</pre>

==== Test point #4 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.445/4.142/5.676/0.868 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 40742ms
rtt min/avg/max/mdev = 2.387/3.969/14.361/0.923 ms, pipe 2, ipg/ewma 4.074/3.681 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 137211ms
rtt min/avg/max/mdev = 3.592/23.098/39.886/4.027 ms, pipe 4, ipg/ewma 13.722/23.760 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.40      73.99   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   535.59     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1316         15079        1597         2344         5314         1857.38      672.09   
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.54.1.7 () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1316         15079        1597         2344         5314         1857.38      672.09   
</pre>

==== Test point #5 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.560/3.960/4.807/0.535 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 40349ms
rtt min/avg/max/mdev = 2.405/3.929/18.161/1.007 ms, pipe 2, ipg/ewma 4.035/3.950 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 137943ms
rtt min/avg/max/mdev = 3.569/23.216/42.977/4.342 ms, pipe 4, ipg/ewma 13.795/24.913 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.37      53.00   
</pre>

Test 5:
<pre>
MNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   524.31     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1339         16299        1594         2341         6097         1902.42      911.95    
</pre>

=== Only lamps on line ===
Vacuum cleaner and hair dryer are devices which are not in use on daily basics and also not for long time. Therefore I repeated tests on last test point with both devices switched off but connected to wiring. Results was similar as with all devices powered on. It looks like saving light bulbs produce serious noise to drop down throughout but adding more noise on same line don't cause too much problems to powerline devices.

==== Test point #5 ====
Test 1:
<pre>
10 packets transmitted, 10 received, 0% packet loss, time 9015ms
rtt min/avg/max/mdev = 3.404/3.590/4.639/0.355 ms
</pre>

Test 2:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 39946ms
rtt min/avg/max/mdev = 2.358/3.893/17.561/1.044 ms, pipe 2, ipg/ewma 3.995/3.745 ms
</pre>

Test 3:
<pre>
10000 packets transmitted, 10000 received, 0% packet loss, time 135802ms
rtt min/avg/max/mdev = 3.426/23.142/42.852/4.098 ms, pipe 4, ipg/ewma 13.581/23.080 ms
</pre>

Test 4:
<pre>
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 87380  16384  16384    60.33      59.04   
</pre>

Test 5:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Local       Local       Remote      Remote      Request Response Elapsed Throughput Throughput 
Send Socket Recv Socket Recv Socket Send Socket Size    Size     Time               Units      
Size        Size        Size        Size        Bytes   Bytes    (sec)                         
Final       Final       Final       Final                                                      
23400       87380       87380       22600       1       1        60.00   525.13     Trans/s    
</pre>

Test 6:
<pre>
OMNI Send|Recv TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to server () port 0 AF_INET
Minimum      Maximum      50th         90th         99th         Mean         Stddev       
Latency      Latency      Percentile   Percentile   Percentile   Latency      Latency      
Microseconds Microseconds Latency      Latency      Latency      Microseconds Microseconds 
                          Microseconds Microseconds Microseconds                           
1305         14034        1592         2343         6077         1904.30      939.65     
</pre>

== Summary ==
Powerline devices from ZyXEL are robust and provide high throughput also under difficult condition. Even with serious noise on line throughput is better that maximum possible via "N" WiFi connection. On other hand there is delay at about 3.5 milliseconds caused by technology itself and limitation on usable sockets (they should be on same line with no "filters" between them). With price at about 46 EUR per pair it is reasonable investment. I recommend this if you would like to use wake on lan or boot via network. Both technologies are not possible via Wifi connection.