Networks are critical infrastructure
A signaling storm in a Norwegian mobile provider coincided with bad flooding in real life, causing very bad public feeling about the reliability of emergency handling.
So a number of operators are focusing on measurements, both in academia and in fixed and mobile broadband.
They measured connectivity between nodes with ping. Mobile operators were less reliable than expected, 1 in 3 connections has a downtime of more than 3 minutes per day. Different operators are good in different places, and have different failure patterns. So connecting to multiple operators improves reliability.
RIPE atlas is a platform for active measurements composed of small probes hosted in currently 1500 places around the world. Data is available to the community in general.
Previously you had to register for RIPE atlas separately, but now there is single sign on, and LIRs are a special, privileged category of users.
User defined measurements based on type, area and targets. Measurements are priced in credits (earned by hosting or sponsoring probes), and are scheduled in a way that avoids overloading the system. Credits and limits can be increased given a good reason (firstname.lastname@example.org).
RIR, BGP GeoIP, etc data from all RIPE databases and histories. Same single sign on as RIPE atlas.
network science at scale
Measuring speed data with full user knowledge in return for non-anonymized data. Currently (after 2 years) 500 TB of open network data from worldwide.
Tools allowing you to run SQL queries against a portion of the dataset.
RIPE Labs, RIPE stat, RIPE Atlas
INRDB Data Store