| 62 | | Similarly we tested frisbee from three FreeBSD nodes using the same technique: |
| 63 | | |
| 64 | | {{{ |
| 65 | | source tb_compat.tcl |
| 66 | | set ns [new Simulator] |
| 67 | | |
| 68 | | set server [$ns node] |
| 69 | | set client1 [$ns node] |
| 70 | | set client2 [$ns node] |
| 71 | | |
| 72 | | tb-set-node-os $server FBSD9-64-STD |
| 73 | | tb-set-node-os $client1 FBSD9-64-STD |
| 74 | | tb-set-node-os $client2 FBSD9-64-STD |
| 75 | | |
| 76 | | set lan0 [$ns make-lan "$server $client1 $client2" 1000Mb 0ms] |
| 77 | | |
| 78 | | $ns rtptoto Static |
| 79 | | $ns run |
| 80 | | }}} |
| 81 | | |
| 82 | | Also, for frisbee we had to manually tell it the bandwidth between our nodes (1Gbit) in order to get it to perform comparably. |
| | 62 | Similarly we tested frisbee from the same three nodes after compiling it for Linux. For frisbee we had to manually tell it the bandwidth between our nodes (1Gbit) in order to get it to perform comparably. This value is actually higher than the perl script that Deterlab uses to start frisbee sets it to. The script in question can be found at [https://github.com/deter-project/testbed/blob/development/tbsetup/frisbeelauncher.in] with lines 75, 76, and 265-272 being of particular interest. |
