Let the Market Drive Deployment: A Strategy for Transitioning to BGP Security
P. Gill, M. Schapira and S. Goldberg
You can download the program demonstrated in the walk through here (5MB .zip). The program has currently only been tested on Windows.
Walk through
When you first load the program it shows a welcome screen and gives you a choice of what simulation result to examine. Here we select simulation 0 which has the 5 content providers and 5 tier 1s as early adopters. This simulation corresponds to the results in section 5.
Once we have loaded the simulation we are given the option to summarize how many ASes have deployed S*BGP in each round.
We can now dig deeper into what is happening in each round. Here we consider the 28th round of the simulation (recall from Section 5, that high degree AS 6939 deploys S*BGP in round 25 causing a ripple effect). We now use the "whoflipped" command to see which ASes deploy S*BGP in round 28.
In the list of ASes that deploy S*BGP in round 28 we see AS 719. To learn more about the impact of AS 719 on paths to its stubs we can use "whosepathchanged". Here we see which paths from Google (AS 15169) to AS 719's stubs changed after AS 719 deployed S*BGP.
By observing the paths that changed when AS 719 deployed S*BGP we can see how this AS gained traffic by flipping. For example, in the highlighted path AS 719 gained traffic from Google to its stub AS 6839 by securing both itself and the stub. In the following, we look at the path from Google to 6839 for each round of the simulation. Google takes the same path until 719 deploys S*BGP to create the secure path.
We observe that Google's path to the stub AS 6839 includes 4 ASes. We may wonder when AS 6667 deployed S*BGP. If it was in a later round this indicates that AS 6667 deploying S*BGP contributed to AS 719's decision to deploy. We consider this by using the "summarize" command for AS 6667. Indeed, we observe that AS 6667 has deployed S*BGP in iteration 27, right before AS 719 decided to deploy.
