We study the detection of the provider-free AS set (PFS), i.e., the set of those Autonomous Systems (ASes) that reach the entire Internet without paying anyone for the traffic delivery. Using trustworthy but non-verifiable sources for sanity checks, we derive the PFS from public datasets of inter-AS economic relationships. Whereas a straightforward method for extracting the PFS performs poorly because the datasets are noisy, we develop a more sophisticated Temporal Cone (TC) algorithm that relies on topological statistics and exploits the temporal diversity of the datasets. The evaluation shows that our TC algorithm detects the PFS with a high accuracy.

The conference will be conducted in English