Data speed tests

[cameo]

I often read reports of data speeds between a PC and various devices
connected to it via WiFi or powerline ethernet but I don't know of any
easy method to test such speeds. It's easy to test the speed between my
ISP and PC using such programs as the Ookla Speedtest, but that's not
much help to check the speed between my PC and an IP camera connected to
it remotely. Can anybody suggest a way to do that?

 

[Char Jackson]

I often read reports of data speeds between a PC and various devices
connected to it via WiFi or powerline ethernet but I don't know of any
easy method to test such speeds. It's easy to test the speed between my
ISP and PC using such programs as the Ookla Speedtest, but that's not
much help to check the speed between my PC and an IP camera connected to
it remotely. Can anybody suggest a way to do that?

As far as I know, the premiere tool is JPerf, which is free.
<http://www.softpedia.com/get/Network-Tools/Network-Testing/JPerf.shtml>
or its close cousin, IPerf.

You run it on two networked systems, where one instance runs as a
server and the other runs as a client. You can control multiple
aspects of the connection (or multiple connections) via the GUI and
see the throughput in real time.

 

[J. P. Gilliver (John)]

I often read reports of data speeds between a PC and various devices
connected to it via WiFi or powerline ethernet but I don't know of any
easy method to test such speeds. It's easy to test the speed between my
ISP and PC using such programs as the Ookla Speedtest, but that's not
much help to check the speed between my PC and an IP camera connected
to it remotely. Can anybody suggest a way to do that?

Doesn't the "Networking" tab in Task Manager give a fair idea? It does
under XP ...

 

[Char Jackson]

Doesn't the "Networking" tab in Task Manager give a fair idea? It does
under XP ...

It shows what's going on, but it doesn't show what the link is capable
of. Its usefulness depends on what you're looking for.

 

[J. P. Gilliver (John)]

I often read reports of data speeds between a PC and various devices
connected to it via WiFi or powerline ethernet but I don't know of any

[]
Doesn't the "Networking" tab in Task Manager give a fair idea? It does
under XP ...

It shows what's going on, but it doesn't show what the link is capable
of. Its usefulness depends on what you're looking for.

In XP, the graph shows (on my machine at the moment, anyway) 0 to 1 per
cent; the table below shows Link Speed of 100 Mbps for the (unconnected)
cable, and 54 for the wireless connection (which is I think the
theoretical maximum for 802.11g). I just loaded a web page, and the
graph axis changed to 0 to 5% for the wireless, with a peak appearing at
about 3% in it; the 54M in the table didn't change.

Or isn't that what you mean by "what the link is capable of"?

 

[Char Jackson]

In message <[email protected]>, cameo <[email protected]>
writes:
I often read reports of data speeds between a PC and various devices
connected to it via WiFi or powerline ethernet but I don't know of any []
Doesn't the "Networking" tab in Task Manager give a fair idea? It does
under XP ...

It shows what's going on, but it doesn't show what the link is capable
of. Its usefulness depends on what you're looking for.

In XP, the graph shows (on my machine at the moment, anyway) 0 to 1 per
cent; the table below shows Link Speed of 100 Mbps for the (unconnected)
cable, and 54 for the wireless connection (which is I think the
theoretical maximum for 802.11g). I just loaded a web page, and the
graph axis changed to 0 to 5% for the wireless, with a peak appearing at
about 3% in it; the 54M in the table didn't change.

Or isn't that what you mean by "what the link is capable of"?

Yes, that isn't what I mean. I meant there are times when you'd like
to see what a network connection is physically capable of, what its
maximum throughput can be. The 100Mbps and 54Mbps figures you cite
above are link speeds, not throughput speeds. In the case of 802.11g,
a link speed of 54Mbps is easily achievable, in which case the maximum
throughput will be about 24-25Mbps, best case. In the case of Fast
Ethernet, the link speed is 100Mbps but the actual max throughout may
be anywhere between 65-95Mbps, depending on many factors. Sometimes a
person might want to know what the network is capable of, and the
Networking tab in Task Manager can't tell you that on its own.

On my Gigabit network, for example, I never see actual throughput
above 600Mbps (usually only 350-400 Mbps) so I got curious. Running
JPerf, I was easily able to sustain 960Mbps, so I know the network
isn't the limiting factor. In my case it's my 'green' 5400 RPM disk
drives, but it was nice to see what the network was capable of.

I have a pair of Netgear 200 Mbps (marketing speak) Powerline
networking units. JPerf tells me they can sustain about 35 Mbps of
throughput. Because of other concurrent network demands, that's not
enough to stream BluRay movies, so I invested in a pair of Zyxel 500
Mbps Powerline networking units. Again, JPerf tells me they can
sustain about 65 Mbps of actual throughput, until the refrigerator
kicks on and drops throughput to about 40 Mbps. If the fridge does
that to the faster units, I wonder what it does to the slower units? I
haven't tested that but I bet it's ugly.

 

[cameo]

As far as I know, the premiere tool is JPerf, which is free.
<http://www.softpedia.com/get/Network-Tools/Network-Testing/JPerf.shtml>
or its close cousin, IPerf.

You run it on two networked systems, where one instance runs as a
server and the other runs as a client. You can control multiple
aspects of the connection (or multiple connections) via the GUI and
see the throughput in real time.

Sorry, I am not comfortable running Java scripts in Windows. I was
looking for something that was compiled into a Windows .exe file.

 

[cameo]

I have a pair of Netgear 200 Mbps (marketing speak) Powerline
networking units. JPerf tells me they can sustain about 35 Mbps of
throughput. Because of other concurrent network demands, that's not
enough to stream BluRay movies, so I invested in a pair of Zyxel 500
Mbps Powerline networking units. Again, JPerf tells me they can
sustain about 65 Mbps of actual throughput, until the refrigerator
kicks on and drops throughput to about 40 Mbps. If the fridge does
that to the faster units, I wonder what it does to the slower units? I
haven't tested that but I bet it's ugly.

I have a similar issue in trying to decide whether WiFi or Homepug AV
powerline gives me a better speed between my PC and IP cams.

 

[Char Jackson]

Sorry, I am not comfortable running Java scripts in Windows. I was
looking for something that was compiled into a Windows .exe file.

I don't understand your decision and how it might apply to this
program, (there's no javascript involved), but I respect your choice.
Good luck.

 

[Paul]

I have a similar issue in trying to decide whether WiFi or Homepug AV
powerline gives me a better speed between my PC and IP cams.

Connect a computer to the other end of the link, and then do your
thruput tests. That will tell you what the performance level of
the Wifi or Homeplug physical layer is.

You can even set up an FTP server and client for
testing if you want (as they're available for a lot of different
OSes). Other options include things like "rcp" as a test case.
Or, a dedicated network performance test suite.

Once you've characterized the link, you can disconnect the computer
and reconnect the IP camera in its place.

Paul

 

[Philip Herlihy]

Connect a computer to the other end of the link, and then do your
thruput tests. That will tell you what the performance level of
the Wifi or Homeplug physical layer is.

You can even set up an FTP server and client for
testing if you want (as they're available for a lot of different
OSes). Other options include things like "rcp" as a test case.
Or, a dedicated network performance test suite.

Once you've characterized the link, you can disconnect the computer
and reconnect the IP camera in its place.

Paul

Try Iperf:

http://en.wikipedia.org/wiki/Iperf

 

[cameo]

I don't understand your decision and how it might apply to this
program, (there's no javascript involved), but I respect your choice.
Good luck.

Sorry, I didn't mean 'Java scripts" per se, but scripts like what that
jperf.bat file contains:
------
start javaw -classpath
jperf.jar;lib\forms-1.1.0.jar;lib\jcommon-1.0.10.jar;lib\jfreechart-1.0.6.jar;lib\swingx-0.9.6.jar
net.nlanr.jperf.JPerf
exit
-------

The the README.txt is not more encouraging either:
--------
********************************************
* INSTALLATION INSTRUCTIONS *
********************************************

* On windows systems,
- Download and install Java JRE 1.5+ on http://java.sun.com
- run the following script : jperf.bat

* On Linux / OS X systems, run the following script :
- The 'java' (JRE 1.5+) executable have to be into the system path
- Don't forget to set execution permissions on the jperf.sh script
(execute 'chmod u+x jperf.sh')
- run the following script : jperf.sh


********************************************
* COMPILATION INSTRUCTIONS *
********************************************

* Go to the 'utilities' directory
* run the following command (Apache ANT has to be installed on the
system) :

ant release

This script will create a JPerf distribution into the 'release' directory.

 

[cameo]

Try Iperf:

http://en.wikipedia.org/wiki/Iperf

What is really the diff between Iperf and Jperf? Only that Iperf
does not use Java?

 

[Char Jackson]

Sorry, I didn't mean 'Java scripts" per se, but scripts like what that
jperf.bat file contains:
------
start javaw -classpath
jperf.jar;lib\forms-1.1.0.jar;lib\jcommon-1.0.10.jar;lib\jfreechart-1.0.6.jar;lib\swingx-0.9.6.jar
net.nlanr.jperf.JPerf
exit
-------

What do you see there that you don't like? What exactly are you
objecting to? Both IPerf and JPerf have been around for years and have
been used by network professionals and curious non-pros without
issues. I'm not sure what it is that's making you uncomfortable.

The the README.txt is not more encouraging either:

Most systems already have Java installed, and if you do you can ignore
the Java installation instructions. Just execute the .bat file and
away you go.

This all sounds like stuff for sys admins, not for simple pedestrian
users like me. :-(

Suit yourself. You're avoiding an excellent tool.

 

[Paul]

What is really the diff between Iperf and Jperf? Only that Iperf does
not use Java?

http://en.wikipedia.org/wiki/Iperf

"GUI

There is a Graphical user interface (GUI) front end available called jperf"

Without the graphical front end, the interface to iperf is command line based.

You start the listener on one computer...

iperf.exe -s -P 2 -i 5 -p 5999 -f k

and the test client on the other computer.

iperf.exe -c s-network1.amcs.tld -P 1 -i 5 -p 5999 -f B -t 60 -T 1 # test client

When you use Jperf, you don't have to use the command line.

http://code.google.com/p/xjperf/

HTH,
Paul

 

[cameo]

http://en.wikipedia.org/wiki/Iperf

"GUI

There is a Graphical user interface (GUI) front end available called jperf"

Without the graphical front end, the interface to iperf is command line
based.

You start the listener on one computer...

iperf.exe -s -P 2 -i 5 -p 5999 -f k

and the test client on the other computer.

iperf.exe -c s-network1.amcs.tld -P 1 -i 5 -p 5999 -f B -t 60 -T 1 #
test client

When you use Jperf, you don't have to use the command line.

http://code.google.com/p/xjperf/

Thanks. Now I see it's not as bad as I thought. However, it still takes
2 PCs and I only have one.

 

[Paul]

Thanks. Now I see it's not as bad as I thought. However, it still takes
2 PCs and I only have one.

Like I explained earlier, you could only test to the IP camera, if the
IP camera supports an interface you can test with. If the IP camera had
a Linux OS inside, it may be possible to load software onto it (the RAM
available inside the camera may not be very large though). If you cannot
access the IP camera in that way, then you're stuck using whatever facility
the camera supports. Some cameras do FTP, but only for motion sensor
"snapshots" the camera has taken. The file size may not be large enough
for a good file transfer rate test.

By replacing the camera with a computer, you have an opportunity to test
the performance level of the Wifi or HomePlug infrastructure. And that
is about as close as you'll get, to finding a bottleneck caused by that.
The IP camera, perhaps it can't go at the full link rate anyway. The
processor in it, being embedded, is not high performance.

Paul

 

[DanS]

I have a similar issue in trying to decide whether WiFi or
Homepug AV powerline gives me a better speed between my PC
and IP cams.

Do the un-named IP cameras have a speed check utility?

How much traffic does the IP cam(s) generate (in mbps)?

Based on Chars numbers above, *his* powerline networking drops to 40mbps when
the refrigerator is running.....54mbps 802.11 *maxes* around 25-30'ish mbps......

.......if each camera uses 3mbps, and let's say 2 camera's...6-7 mbps. Either link alone
is enough to handle both cameras.

But, what else do you do on either of those networks?

Just simply knowing how fast a certain leg is capable of doens't automatically translate
into it being the best coice.

 

[cameo]

Do the un-named IP cameras have a speed check utility?

No, they don't.

How much traffic does the IP cam(s) generate (in mbps)?

Based on Chars numbers above, *his* powerline networking drops to 40mbps when
the refrigerator is running.....54mbps 802.11 *maxes* around 25-30'ish mbps......

......if each camera uses 3mbps, and let's say 2 camera's...6-7 mbps. Either link alone
is enough to handle both cameras.

Good news! I just got a pair of new Trendnet HomePlug AV 200 adapters
(TPL-303E2K) that are the upgraded replacements to my older TPL
-302E units and completely compatible. Even though I didn't plan to
install them with the setup CD that came with them because they could
pair with the existing ones without it, I discovered that the utility on
the CD was actually measuring the speed of each unit to the main one
attached to the router. It's amazing what I discovered with that
utility. Mainly how the speed could vary just which outlet you stick the
adapter into. Even between two outlets on the same phase, circuit and
room! So with this I was able to find the fastest configuration for may
cams, in the 70-80 mbps range. That sure is faster than running the cams
with their own 54G WiFi.

But, what else do you do on either of those networks?

Nothing else really. I assume the cams don't need more than 1 Mbps each,
due to the mp4 compression. I think the limiting factor, if anything, is
going to be the Comcast upload speed that for me right now is around 3 Mbps.

Just simply knowing how fast a certain leg is capable of doens't automatically translate
into it being the best coice.

Only with everything else being equal.