Results

The results of December 15, 1999 are shown here by including some of the Web available HTML tables. Please note that the selected day is typical: the results do not vary much during time, but off course there is a difference between the weekends and the working days. We do not show here the long term statistics which are also available (page ; subsection 2.2), because there are not much short time fluctuations in the network here. Figure 2 (figure 3) shows the HTML table of the the minimum (average) roundtrip values [ms] between the VU site and the other sites. Figure 4 shows the throughput values [Mbit/s] for those connections. The tables are fragments of the corresponding Web pages. In both tables the background of the table cell reflects the value of the roundtrip or throughput value: black for the minimum and white for the maximum. These pseudo gray values may help to find tendencies more easily. Please note that normally pseudo colors are used, but for the sake of reproduction we limited ourself to gray values.

  

Figure: The HTML table with the minimum roundtrip values [ms] between the VU site (table 3) and the other sites. It is a fragment of the corresponding Web page. The background of the roundtrip table cell is a reflection of its intensity. Show figure in real size.

  

Figure: The HTML table with the average roundtrip values [ms] between the VU site (table 3) and the other sites. It is a fragment of the corresponding Web page. The background of the roundtrip table cell is a reflection of its intensity. Here the background colors are logarithmic scaled; linear is the default. Show figure in real size.

  

Figure: The HTML table with the throughput values [Mbit/s] between the VU site (table 3) and the other sites. It is a fragment of the corresponding Web page. The background of the throughput table cell is a reflection of its intensity. Show figure in real size.

From the figures 2, 3 and 4 the following observations can be made:

• The ATM minimum roundtrip values give a reasonable reflection of the geographical distance. The reason therefore, is that all ATM routes require one hop.
• There is not much difference in minimum roundtrip values between the ATM connections (columns entitled ``ATM'') and the ``usual'' Internet route (columns entitled ``Inet''), however, the differences and variations are larger for the average roundtrip values, especially during daytime. This suggests some congestion at the ``Inet'' route.
• The ATM throughput values are constant and close to the expected maximum of about 5 Mbit/s which follows a.o. from the # reserved cells. We do not know why the throughput of the VU - RUL connection is somewhat lower. The DAS sites are described in table 3.
• The variations in the ``usual'' Internet throughput values, reflect the load of the Internet during the day. But the ATM throughput values remain rather constant during the day. This probably is an indication that the capacity of the ATM connections is not yet fully used. These observations are typical. Note that for distributed computing short roundtrip times and constant bandwidth are important for message passing.