You're viewing an archived page. It is no longer being updated.
The Test Traffic Measurement Service (TTM) was shut down on 1 July 2014. This information is available for historical reference.
The test boxes that are on the net are named ttXY, where XY is a two-digit number. To find out which test-box is at your site, simply look on the front panel of the test-box. The test-boxes tt01-tt04 are at the RIPE-office in Amsterdam.
On the real page, you'll see a matrix that looks like this (with the hyperlinks actually working):
From: |
To: |
||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
tt01 |
XX Internet, Somewhere | Overview | tt02 | tt03 | tt04 | tt07 | tt08 | tt12 | tt13 | tt14 | tt15 | tt16 | |
tt02 |
XX Internet, Somewhere | Overview | tt01 | tt03 | tt04 | tt07 | tt08 | tt12 | tt13 | tt14 | tt15 | tt16 | |
tt03 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt04 | tt07 | tt08 | tt12 | tt13 | tt14 | tt15 | tt16 | |
tt04 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt07 | tt08 | tt12 | tt13 | tt14 | tt15 | tt16 | |
tt07 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt08 | tt12 | tt13 | tt14 | tt15 | tt16 | |
tt08 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt07 | tt12 | tt13 | tt14 | tt15 | tt16 | |
tt12 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt07 | tt08 | tt13 | tt14 | tt15 | tt16 | |
tt13 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt07 | tt08 | tt12 | tt14 | tt15 | tt16 | |
tt14 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt07 | tt08 | tt12 | tt13 | tt15 | tt16 | |
tt15 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt07 | tt08 | tt12 | tt13 | tt14 | tt16 | |
tt16 |
XX Internet, Somewhere | Overview | tt01 | tt02 | tt03 | tt04 | tt07 | tt08 | tt12 | tt13 | tt14 | tt15 |
These boxes were used to measure the network delays between each-other. By clicking on any element in the To: colomns of the matrix below, you can see the delays from a particular test-box to another test-box for a 24h, 7 day and 30 day period. For more details about these plots, click here. Clicking on the name of the sending test-box (the From:) colomn, returns the GPS Receiver conditions for this test-box. The overview column gives you a quick overview of all plots for the box.
The overview plots page looks something like:
TTbox | INcoming traffic | OUTgoing traffic |
---|---|---|
tt02 | ||
tt03 | ||
tt04 |
The two columns show a reduced version of the 7 day plots for INcoming (tt02 to tt01, tt03 to tt01, ...) and outgoing (tt01 to tt02, tt01 to tt03, ...) traffic. By clicking on a particular plot, a new window is opened showing the 24h, 7 day and 30 day plots for that connection. This can be used as a quick way to spot problems.
On these pages, you'll find 3 sets of 4 plots. A set typically looks like:
At the very top, a header box lists the path (source & destination box) and time frame (start and end date) of the plotted measurements. On the right a statistics pad summarizes the results.
The 4 plots themselves show the following:
The plots are generated for the last 24 hours, the last 7 days and the last 30 days. They are updated once a day.
The overview plots page looks something like:
_
TTbox | INcoming traffic | OUTgoing traffic |
---|---|---|
tt02 | ||
tt03 | ||
tt49 |
The two columns show a reduced version of the 7 day jitter plots for INcoming (tt02 to tt01, tt03 to tt01, ...) and outgoing (tt01 to tt02, tt01 to tt03, ...) traffic. By clicking on a particular plot, a new window is opened showing the 24h and 7 day day plots for that connection. This overview can be used to get a quick view of the jitter patterns on all connections of a box and easily zoom in on the ones of interest.
On these pages, you'll find 3 sets of 4 plots. A set typically looks like:
At the very top, a header box lists the path (source & destination box) and time frame (start and end date) of the plotted measurements. On the right a statistics pad summarizes the results.
The 4 plots themselves show the following:
The accuracy of the local clock, and thus the accuracy of these measurements, depends on the number of satellites and the signal-to-noise ratio that the GPS receiver sees. At the very minimum, it should have seen the signals from at least 3 satellites for a few hours once at its present location, in order to determine its position. As long as the receiver is not moved after that, it should see the signal from at least 1 satellite all the time in order to keep its clock synchronized. 1 satellite is the absolute minimum, if it sees more satellites it will improve its position determination (which, in turn, improves the time keeping) as well as cross-check the information from the various satellites.
When the receiver has not determined its current position yet, or when it has lost contact with all satellites, then the local clock may slowly start to drift away from the correct time. After a while, NTP will detect this and start to return invalid time-stamps. However, before this happens, the delays may appear to slowly change over time even though nothing changes in the networks.
However, as long as the test-box sees at least 1 satellite (and the antenne is not moved), then this will not happen. This is why the test-boxes record the number of satellites that they see every 64 seconds. This information is binned and clicking on the name of the machine gives you a histogram with the number of satellites seen at 64 s intervals over a 24h, 7 day and 30 day period. Here is an example of such a plot:
As you can see, most of the entries are in the (blue) bins around 3 and 4, which means that the receiver saw 3 or 4 satellites at that time. As we explained before, this is sufficient to keep the clocks synchronized. Here is another example:
In the bottom plot, you see that there are lots of entries in the (white) bin around 0. This means that most of the time, the receiver did not see any satellites.
When you see a large number of entries in the bin around 0, it is recommended to find a better spot for the antenna. For this receiver, this happened during the last week: the 30 days plots shows lots of entries in the 0th bin but this disappeared over the last week.
Finally, number of satellites is an important factor to determine the quality of the local time, but it does not contain all information. We are working on a better way to express the clock quality.