Introduction

In this version of the package the network performance was monitored between the sites described in section 3. In the comparison of the results we will focus to the connections from IPP and ZAM to FOM, SARA and UU. The topology of the involved providers is shown in figure 3.

  

Figure: Topology involved providers.

Its is expected that a bandwidth of 10 Mbit/s is required for the connection with the FZJ. In our analysis we will use this values as a target.

During the month January 2000, the network seems to be rather stable, although the throughput values during daytime in the first week of January were higher than in the other weeks. The reason probably is that in the first week of January many people still did have holidays. Nevertheless, besides the typical daily data we also now have the opportunity to present mean data.

As typical daily data we selected the two days from Wednesday 26/01/2000 23:00 back in history until and included Tuesday 25/01/2000 00:00. These days show a rather typical behavior, however, incidentally during daytime the throughput values between Germany and the Netherlands were still somewhat worse than shown in these pages.

For the mean data the workday mean values averaged during January 2000 were selected. Please note that the mean values are calculated at the periodic measurement times, such that still a distribution across the day is obtained. See section 2, page .

Please note that there are also mean values available for the days of the week, averaged at an quarterly basis. But we feel that after one month the statistic is still insufficient: the average at each measurement time and at each day of the week is only based at four or five individual throughput values.

We do not present roundtrip results here. the reason is that we found no, or hardly any, unusual observations: no unusual long roundtrip times and no unusual package lost.

In the following sub sections we will make several comparisons between the performances of the sites.