Zeph is a reinforcement learning based algorithm for selecting prefixes to probe based on previous measurements in order to maximize the number of nodes and links discovered. Zeph can be used on top of the Iris platform.
Zeph has a command line interface to configure and run the algorithm.
First, install the Zeph package:
pip install dioptra-zeph
Zeph takes as input a list of /24 (IPv4) or /64 (IPv6) prefixes:
# prefixes.txt
8.8.8.0/24
2001:4860:4860::/64To start a measurement from scratch:
zeph prefixes.txtTo start from a previous measurement:
zeph prefixes.txt UUIDZeph relies on iris-client and pych-client for communicating with Iris and ClickHouse. See their respective documentation to know how to specify the credentials.
You can create an exhaustive list of /24 prefixes from a BGP RIB dump:
pyasn_util_download.py --latest
# Connecting to ftp://archive.routeviews.org
# Finding most recent archive in /bgpdata/2022.05/RIBS ...
# Downloading ftp://archive.routeviews.org//bgpdata/2022.05/RIBS/rib.20220524.1000.bz2
# 100%, 659KB/s
# Download complete.
zeph-bgp-convert --print-progress rib.20220524.1000.bz2 prefixes.txt@article{10.1145/3523230.3523232,
author = {Gouel, Matthieu and Vermeulen, Kevin and Mouchet, Maxime and Rohrer, Justin P. and Fourmaux, Olivier and Friedman, Timur},
title = {Zeph & Iris Map the Internet: A Resilient Reinforcement Learning Approach to Distributed IP Route Tracing},
year = {2022},
issue_date = {January 2022},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
volume = {52},
number = {1},
issn = {0146-4833},
url = {https://doi.org/10.1145/3523230.3523232},
doi = {10.1145/3523230.3523232},
journal = {SIGCOMM Comput. Commun. Rev.},
month = {mar},
pages = {2–9},
numpages = {8},
keywords = {active internet measurements, internet topology}
}Iris is developed and maintained by the Dioptra group at Sorbonne Université in Paris, France.