CAIDA / submarine-cable-impact-analysis-public

This repository contains tools implemented for the PAM 2020 paper "Unintended consequences: Effects of submarine cable deployment on Internet routing" to collect and analyze data depicting the impact of the South-Atlantic Cable System (SACS) launch on Internet routing. This codebase can be extended to other use-cases of cable launches, failures, etc.

Note: This repository is still being updated.

PAM Paper: Unintended consequences: Effects of submarine cable deployment on Internet routing (https://www.caida.org/publications/papers/2020/unintended_consequences/)

submarine-cable-impact-analysis

Our methodology can be divided into 4 steps. Each of them corresponds to the first 4 directories of this repository:

• Step1-collect-candidate-paths-crossing-the-cable
• Step2-identify-router-IPs-on-both-sides
• Step3-search-for-comparable-historical-traceroutes
• Step4-annotate-collected-paths

We added in Step5-data-analysis the scripts used for data analysis

Python packages to install for running the scripts smoothly

E.g. pip install --target=/usr/local/bin pyasn

• Wandio
• urllib2 (Python2) or urllib (Python3)
• IPy (https://pypi.org/project/IPy/)
• pyasn (https://pypi.org/project/pyasn/)
• dnspython (https://pypi.org/project/dnspython/)
• docopt (https://pypi.org/project/docopt/)
• pycountry-convert (https://pypi.org/project/pycountry-convert/)

Brief description of the folders' content

Step 1: Collect candidate IP paths that could have crossed the cable.

This folder contains the scripts for 2 possible ways for achieving this goal (Method1-Use-Henya-and-Ark-VPs and Method2-Use-RIPE-Atlas-VPs). The first method can be used for Ark VPs, while the second can be used for RIPE Atlas probe.

Step 2: Identify router IP interfaces on both sides of the cable.

To identify the router IP interfaces, you can run the script 1_identify_router_IPs_from_ITDK_data_v2.py. The links IPs are listed in Final_list_IPs_RA.txt, and Final_list_IPs_RB.txt

Step 3: Search for comparable historical traceroutes.

Step 4: Annotate collected paths.

Step 5: Data analysis.

You can use the script 1_Finding_Consistent_AS_pairs_transiting_cable_operating_AS.py to find consistent AS pairs transiting the AS operating the cable, and 2_Computing_lengths_AS_paths_between_consistent_AS_pairs.py to compute the length of AS paths between the same AS, destination prefixes pre & post SACS.

Access to the data

Collected and analyzed traceroutes and other datasets generated during the analysis are available here: http://data.caida.org/datasets/supplement/2020-unintended-consequences/