Initial Setup

After some inquiries at the TEN-155 community, it appears that a good starting setup would be between DFN, SURFnet and JANET. One of the reasons is that these are the three points around the current ring UK - NL - DE - FR where most of the traffic flows. Traffic between JANET and DFN goes via France. This setup is drawn in figure 11.

  

Figure: Proposed start setup for net performance tests at the TEN-155 network. Note that the route between JANET and DFN goes via France.

So we would suggest to start the TEN-155 performance tests with three workstations at or near the backbones of the NRN's. Because the # sites is limited we would suggest to test the full matrix at an hourly period. This implies two roundtrip and two roundtrip measurements for each site. As explained in section 3, the roundtrip tests are executed simultaneously, so using the parameters of subsection 3.1 this means an engagement of each site with the package of 60 seconds. With the suggested test period of an hour, this is a relative load of 1.7 % which is rather moderate. Considering that only default socket parameters are used at the throughput measurements we do not expect congestion problems. The tests at the DAS Cluster (subsection 3.2) are running longer than a year and the tests at the IPP - FOM setup for about half a year without any problems.