Introduction

The MANIC project -- Measurement and Analysis of Interdomain Congestion -- has developed a prototype system to monitor interdomain links and their congestion state, in order to provide empirical grounding to debates related to interdomain congestion. Its dataset is collected using the Time-Sequence Latency Probing (TSLP) method from Ark Vantage points, after running the bdrmap algorithm to automatically infer the IP links (the near IP and far IP addresses) interconnecting neighboring Autonomous Systems (ASNs). We will soon integrate MANIC with CAIDA's AS Rank platform to allow more complex queries of the Internet's interconnection infrastructure.

Time-Sequence Latency Probing (TSLP) method

Consider two neighboring networks ISP A and ISP B, interconnected by interdomain link A-B of interest. The corresponding physical link interconnects the edge ("near") router located at the border of ISP A with the edge ("far") router at the border of ISP B (see figure). The TSLP process probes a destination "behind" the interdomain link with TTL set to expire respectively at the near and the far ends of the link. Repeating this operation periodically yields an RTT timeseries for both sides of the link. When the buffer occupancy at the interdomain link increases, RTTs to the far end increase, which indicates evidence of congestion on the interdomain link. For more details, see our recent publication at ACM SIGCOMM 2018.