|
|
|
|
Speed test for Internet access Frequently asked questions (FAQ) |
1 How to proceed during measurements
1.1 My
provider doesn't offer the speed test facility. Can I test the speed anyway?
1.3 I am
a DSL customer. What shall I do if the measured results are not up to my
expectations?
1.4 Is
my DSL connection limited?
1.7 Why
do I measure a download rate which is far too high?
1.8 Why
do my values get displayed in red if I measure unrealistic high values?
1.9 Why
do I get different performance values if I start the test applet on different
test servers?
1.10 When
using GPRS/UMTS why do
I get unrealistic high response times of under 50 milliseconds?
1.11 Which
parameters are taken account of during the speed test?
2 Problems
while carrying out the test
2.3 How
can I activate cookies?
2.4 How
can I activate JavaScript?
2.5 After the
measurement operation, I get a message: "Unable to save results";
why is that?
2.6 Why
does it take more than 20 seconds till the speed test starts?
3 Analysis
of the tests (all your results)
3.1 Why
do I get a red warning "work with a LAN cable"?
3.3 Why
do I get 127.0.0.1 or 0.0.0.0 displayed
as internal IP address?
3.6 How can I
delete old measurements from the statistics?
All providers can carry out the speed test if a license fee is payed to cnlab. If your provider has not taken up that offer and does not have its own test server, you can select another one. Then you will not only measure the performance of the network of your provider but also the performance of the connections between your provider and the test server that was linked to.
The performance promised by the Internet Service Providers (ISP) refers to the LAN cable connection to your modem. If your computer is not directly connected (via a LAN cable) to your modem, but via wireless LAN (WLAN), USB or power line, the data rate may be limited.
The maximum bandwidth of wireless LAN connections depends on the wireless standard, the distance and barriers (walls) between your computer and the wireless access point (router) as well as by the load of your wireless devices and other troublemakers. The main interference in the WLAN frequency range are cordless phones, baby monitors (or a variety of video transmission systems), microwave
ovens or devices that communicate via Bluetooth.
You also share the
overall available bandwidth of the WLAN with neighbors, working with the same wireless channels. With programs such as InSSIDer you can get an idea about the signal strength and channel selection of other WLANs.
Note that even devices with a specified data rate of 300Mbps only achieve data rates exceeding 100Mbps under optimal conditions.
To check the connection
speed, proceed as follows:
1.
Note
the download data transfer rate of your DSL offer:
Data transfer rate of the
offer = ………………. Kbits/sec
2.
Use the option for checking the
transfer rate of your telephone line which is to be found on the website
of Swisscom or the respective offers of your provider (for example Tele2, Solnet, Green, Orange, Tiscali/VTX); in that way you
find out about te maximum data tranfer rate which is possible under
consideration of the lenght of your telephone line:
Data transfer rate
for the connection = ………………. Kbits/sec
3.
Your telephone line should be able
to reach a "target data transfer rate" which corresponds to the
smaller value or both of the values under section 1 and 2 above. Note this
"target data transfer rate":
Target data transfer rate = Minimum
taken from transfer rate of offer and transfer rate of line (from point 1 and
2) = ………………. kBit/s
4.
Go to http://speedtest.cnlab.ch and select the
test server of your own Internet Service Provider or a test server close to
your home and carry out a number of measurements. Make sure that while carrying
out the measurements there are no programs using up considerable resources
running on your computer. Also make sure that at the time of measurement, your
internet connection is not supporting the load of other stations (if you are
subscribed to Bluewin TV, this should not run while you carry out the
measurements).
If
the measured value exceeds 90% of the target transfer rate then no further
steps need to be taken.
In that case your internet
connection is in compliance with the expectations.
5.
If the measurement definitively
gives worse results, then carry out further measurements during periods where
there is a low load on resources, (at night for example). If even during such
periods you are far from reaching the target data transfer rate, then it could
be that the lack of performance is caused by your local network connection
appliances (e.g. WLAN, powerline, router).
Connect your computer directly to the power unit or to the DSL modem
respectively and repeat the measures. (Note that when carrying out measurement
via WLAN it is possible to get a substantially lower data transfer rate than if
you measure directly at the power unit. The reason for that is that wireless
transmission is more exposed to disturbances and because typically the data
transfer rate is lower than in the case of transmission via cable).
If
the value you measure now exceeds 90% of the target data transfer rate, look
for defaults at the level of your local power units as your Internet connection
is satisfactory.
6.
If the target data transfer rate
is still not reached although your computer is connected directly to the modem,
it could be that the problem is at the level of your computer. If possible,
repeat the measures with another computer or notebook.
If
now measurements on another computer exceeds 90% of the target data rate, look
for defaults at the level of your computer.
If however, the target data transfer rate is far from being attained
although measurements have been carried out on various dates of the week, then
use the form "Reporting a problem" (at the bottom of the speed test
page) to inform your Internet service provider or consult a network expert.
If you have a long telephone line or one of poor quality, only rather limited data transfer rates are possible; The links below will serve to check whether your DSL connection is subject to limitations or not. In general the limitations should not depend on the provider..
· Swisscom/Bluewin: http://www.swisscom.ch/res/internet/dsl/index.htm?languageId=de
· Green: http://www.green.ch/greendsl/pruefen_DSL.asp
· Solnet: http://www.solnet.ch/produkte/zugang/DSL/linechecker.html
· Sunrise: http://www.sunrise.ch/privatkunden/angebote/iminternetsurfen/dsl.htm
· Tele2: http://www.tele2.ch/de/internet-faq.html
· Tiscali/VTX: http://surf.vtx.ch/Check_surf.aspx
To check the connection speed, proceed as follows:
1.
Note the download data transfer
rate of your Cablenet offer:
Download data tranfer
rate = ………………. kBit/sec
2.
Go to http://speedtest.cnlab.ch and select the
test server of your own Internet Service Provider or a test server close to
your home and carry out a number of measurements. Make sure that while carrying
out the measurements there are no programs using up considerable resources
running on your computer. Also make sure that at the time of measurement, your
internet connection is not supporting the load of other stations in your home.
If the measured value exceeds
90% of the target transfer rate then no further steps need to be taken. In that
case your internet connection is in compliance with the expectations.
3.
If the measurement gives values
which fall clearyly below the target value, you should carry out further
measurements and if possible do so outside of periods where there is a high
load (for example between 11 p.m and 7 a.m). For that purpose you can use the
function "24 measurements in 24 hours" and display the results of all
measurements using the option "more details".
If
even during the periods of low load you cannot reach at least 90% of the data
transfer rate stipulated by the offer of your service provider, it is likely
that the network of your computer environment is overloaded. In some cases, it helps if you restart your
modem (version with cable) by switching it off and putting it on again after a
wait of two minutes. If two or three
restart attemps bring no remedy, it is probably necessary to rebuild the
infrastructure of your computer environment.
4.
If even during periods of low load
the data transfer rate stipulated for your offer is not reached, it could be
that there is a problem with your power connection units (for example WLAN,
Powerline). Connect your computer directly to the the power unit or to the
modem (version with cable) and repeat the measurements. (Note that when
carrying out measurement via WLAN it is possible to get a substantially lower
data transfer rate than if you measure directly at the power unit. The reason
for that is that wireless transmission is more exposed to disturbances and
because typically the data transfer rate is lower than in the case of
transmission via cable).
If
the value you measure now reaches 90% of the target data transfer rate that
means that your Internet connection fullfills the requirements. Look for
defaults at the level of your local power units as your Internet connection is
satisfactory.
5.
If the target data transfer rate
is still not reached even during the periods of low load and although your computer
is connected directly to the modem, it could be that the problem is at the
level of your computer. If possible, repeat the measures with another computer
or notebook.
If
now measurements on another computer exceed 90% of the target data rate, look
for defaults at the level of your computer.
If however, the target data transfer rate is far from being attained
although measurements have been carried out with other computers directly
connected to the modem and on various dates of the week, then use the form
"Reporting a problem" (at the bottom of the speed test page) to
inform your Internet service provider or consult a network expert.
Depending on the circumstances, data transfer rates are indicated in "bits per second" (for instance when indicating connection speeds) or in "bytes per second" (for example in the case of downloads or for file transfer programs). A byte contains 8 bits. Therefore the data transfer rate indicated in "bits per second" should be 8 times higher than that indicated as "bytes per second". Therefore the 250 kb/sec rate shown in the download window is equivalent to 2000 kb/sec.
Too high data transfer rates typically have two reasons:
1. Values which exceed those expected by up to 20% : Many Internet service providers have configured their systems in such a way that under optimal circumstances 120% of the proposed data transfer rate can be attained. Therefore it may be that if for instance your offer provides for 5000 kbits/sec you reach values of up to 6000 kbits/sec.
2. If you have values which exceed those expected by more than 20%: if you are using a proxy server it can happen that test files on the proxy server get into intermediate storage and are thereafter transferred with LAN speed (the latter will then be erroneously displayed as data transfer rate). It might be that you have a HTTP proxy software installed. Such a program is for instance contained in the Gdata anti-virus kit (OutbreakShield technology). To get correct results with the speed test, turn off the HTTP proxy server or use another computer for the measurements.
The maximum data transfer rate for upload or download per tester on the test server is limited so as to avoid overload of the server. The values are displayed in "red" in the case where your data transfer rate is artificially limited owing to that limitation procedure. The reason can be that you are having a very fast connection (for instance in a corporate or university environment) or that a local local proxy server (which may be part of your virus scanner or security software) carries out an intermediate storage and gives the impression of a high data transfer rate.
The speed test functions in such a way that you check the connection between your own computer and the test server from which the applet has been loaded. If you load the applet from the test server of your own provider, you are mainly testing the connection network of your provider. If you load the the applet from other test servers you also test the connections between the providers.
On the mobile network the response times are usually longer than 50 milliseconds. If however you use special software to optimize your data transfer (e. g. SpeedOptimizer) it could happen that the test is falsified and gives unrealistic low values. In that case, deactivate that software and repeat your measurements.
In order to give best results, in the course of the tests the entries of the tester and also the results and data from the computer are captured and saved:
· Entries of the testers: Information on the connection (Type of connection, speed or offer respectively), Indications regarding the location (country, postal code, location)
· Results of measurements: time of measurement, data transfer rate for download, data transfer rate for upload and time of response
· Server to which the measurement was carried out
· Local IP address and external IP address
· Technical data regarding the system of the tester (e.g. settings of the browser and the computer, TCP settings)
Every day there are about 15 000 tests carried out from about 3000 computers. After advertisements in the media or at the time where offers get changed the frequency of measurements increases by 10. When doing statistical evaluations the measurements are grouped by computer. Therefore the results from testers who carry out a large number of tests don't weigh more than the results of testers who have carried out just a few measurements. Statistical evaluations are based on the rule that at least 100 values must be available before concluding about an effect observed.
For two reasons the applet measurements give rather
pessimistic values: people usually carry out tests when they feel that
performance is inadequate and also people tend to make the tests during periods
of high load.
Following information given by Cablecom for the year 2007, it was found out
that in regions where performance is recognised to be low there were 12 series
of measurements carried out per 100 clients (or 27 measurements per client)
whereas there were only 8 series of measurements per 100 clients (or 59
measurements per 100 clients respectively) in regions where performance is
good. From that we see that in networks where performance is poor more clients
carry out measurements and do series of measurements with a larger number of
measures.
The speed test is working with what is called a JavaApplet. JavaApplets are programs which are called by the web server and executed in the browser. Go to http://www.java.com/de/download/installed.jsp?detect=jre&try=1 to check if the latest Java version is installed on your system. The Java test is also available under http://www.heise.de/security/dienste/browsercheck/.
With many browsers the JavaApplets are not supported under the default settings. If that is the case, you will have to load the current Java version by going to http://www.java.com/de/ to be able to carry out the speed test. (After you have installed Java, the browser must be restarted in any case).
It can also be that Java has been installed but is not currently activated in your browser. Proceed as follows to activate Java:
· Internet Explorer: Tools-> Manage Add ons -> Enable or Disable add-ons
·
Firefox: Tools -> Options ->
Content
select the option "Enable Java"
· Safari: Safari -> Tools -> Security -> Enable Java
It can happen that the speed test is gets influenced by special programs, virus scanners, firewalls or similar devices and so because JavaApplets were not correctly executed. You will for instance get an error message if an incompatible Java version had been installed or if an older Java version was not updated correctly.
Go to http://www.java.com/de/download/installed.jsp?detect=jre&try=1
to check if Java has been correctly installed in the latest version.
If the speed test still doesn't function correctly you may need to suppress the
older Java versions manually and reinstall Java. In any case, the browser must
be restarted after a change to save the changes.
Cookies are used to group measurements and to save the settings and results of the testers. If cookies are not supported by your system, you can neverhteless carry out measurements. However, it that case it will not be possible to call up the results of those measurements later. To activate cookies proceed as follows:
·
Internet Explorer: Tools ->
Internet Options -> Security
select the option Medium-high.
·
Firefox: Tools -> Options ->
Privacy
select the option "Accept cookies“
· Safari: Safari -> Tools -> Security -> Accept cookies
Go to http://www.heise.de/security/dienste/browsercheck/ to check if cookies are supported.
JavaScript is used to simplify the operation and entry of user data. If JavaScript is not supported on your system the speed test cannot work.
To activate JavaScript proceed as follows:
·
Internet Explorer: Tools ->
Internet Options -> Security
select the option Medium-high.
·
Firefox: Tools -> Options ->
Content
select the option "Enable JavaScript"
· Safari: Safari -> Tools -> Security -> Enable JavaScript
Go to http://www.heise.de/security/dienste/browsercheck/ to check if JavaScript is active.
The results of your measurement rows are saved on a central server. It can happen that the central server is temporarily unavailable. However, your measurement values don't get lost. Repeat the measures on another day. It is very unlikely that the server will be unavailable for several days.
Delays when starting the speed test (while displaying "Please wait – performance test gets started") can occur for two reasons:
In general, before first using Java, the Java runtime environment must be started. Depending on the speed of your PC this process can take up to 20 seconds.
The speed test will contact the test server via one of the ports within the range of the ports 8008 to 8027. This is done to make sure that there aren't too many measurements taking place at the same time. In case where a firewall and/or virus scanner inhibits the connection to these ports, the test server will be called up via the standard port 80 after about 10 seconds (this situation can for instance be observed when a Bluewin firewall is activated with the settings "High" or "Medium").
In the case of high-speed offers local connection devices (e.g. Wireless LAN Access Points or Powerline connection units) are ofting responsible for reduced performance. For that reason, testers are informed that it is best to connect the computer directly to the connection unit of the Internet service provider (DSL modem/router, Cablenet modem).
While the speed test is carried out information concerning the IP
address under which you registered under the test server is captured. This
address will be referred to as your "public address". Furthermore the
current IP address of your computer that is your "internal address"
will be identified. The two addresses will be compared.
In case where your internal IP address is the same as your public IP address,
it will generally be assumed that you do not use a NAT router and you will get
the message: "Your computer is
appearantly directly reachable via the Internet and therefore exposed to
potential danger".
In case where these IP addresses differ, it will be assumed that you use a NAT
router/Firewall and you will get the message: "It seems that your
computer cannot be directly reached via the Internet. It seems that you use a
NAT router or a firewall".
If 0.0.0.0 is displayed that means that the
internal IP address is not correctly recognised. If 127.0.0.1 is displayed that
can be owing to the protective functions of special programs (for instance
firewall/anti-virus software with proxy functions or due to certain SpeedOptimizer
programs).
You don't have to do anything as the incorrect recognition of the internal IP
address does not affect the measurement results.
The Receive Window Size Rwin describes the maximum data volume that the sender may send out before getting an aknowledgement from the receipient. If the delay time (Round Trip Time, RTT) within the network is so high that the sender has send off the maximum data volume before the aknowledgement is received, then the data transfer rate is limited by the control mechanism of the TCP process. If you have a round trip time of 100 milliseconds and a Receive Window Size of 17200 Bytes, the download data transfer rate will be limited to 1’372k bits/sec although the actual connection speed is much higher. If the Receive Window Size is limited to the maximum of 65535 bytes then the data transfer rate will be limited to 5 242k bits/sec.
Licensees of the cnlab speed test offer the possibility to send the results of your measurements to the customer support service via a Troubleticket. If however you carry out the test on the connection of a provider who doesn't support the cnlab speed test, you will still be able to test the quality of the connection but not to report the problems directly to the provider by using a Troubleticket.
Your measurements will be grouped as "measurement rows". You can separate old measurements from new ones using the function "start new measurement row". The results of the old measurement rows will still be available at the bottom on the right on the listing of the measurement rows.
In case where you want to delete old measurement rows or no longer access the old measurement results, you will have to delete the cookies of the corresponding test server on your browser. The cookies can be deleted as follows:
· Internet Explorer: Tools -> Internet Options -> General -> Delete -> Delete cookies
· Firefox: Tools -> Options -> Privacy -> Show cookies -> Remove cookie
·
Safari: Safari -> Tools ->
Privacy -> Show cookies -> Remove