We provide experiments with runner scripts to reproduce our results claimed in our ACM SIGCOMM2021 paper.
Experiments were tested in a test VM instantiated by the Vagrantfile in the repo root. For the interoperability an additional VM is required (details).
Please refer to the howto
We present the following experiments to demonstrate SAGE capabilities and to reproduce main results of the paper:
Name | Scripted/Manual | Execution Time |
---|---|---|
Validation of checks on IGMP | Scripted | 2 min |
Interoperation with existing tools ICMP | Manual | depends (>15 mins) |
Demonstration on human involvement for bad message | Scripted | 2 min |
Show the manual work on NTP | Mixed | depends (>10 mins) |
Reproduce Figure 5 | Mixed | 27-30 min |
Reproduce Figure 6 | Mixed | 78 min |
Additonal experiments (e.g., BFD processing) are available at SAGE tests.
Navigate to a experiment folder. Test steps are described in the README. Please, consult with the experiment README first.
Experiments rely on a runner script run.sh
that prints either OK
or FAIL
depending on the experiment outcome.
An example of experiment execution:
$ cd validation_of_checks_igmp
$ ./run.sh
[IGMP] Resetting..
[IGMP] Executing..
[IGMP] Checking results..
[IGMP] Check: Logical Forms OK
[IGMP] Check: Headers OK
[IGMP] OK