Risk management is central to modern operational cybersecurity practice, so naturally there's great interest in reported domain risk scores. DomainTools currently computes proprietary risk scores for all effective 2nd-level domains, giving each domain a score from 0 to 100. Those risk scores can help users to decide if a given domain is safe or may be too risky.
While domain-by-domain risk scores are useful in their own right, could risk scoring could be extended to a larger aggregate, such as entire autonomous systems? Perhaps there are "safer" autonomous systems and "riskier" autonomous systems, just as there are safer and riskier registered domains?
We computed aggregate risk score for major autonomous systems (ASNs), looking specifically at registered domains that begin with a digit. Our findings from this work include:
Details about the approach employed and a copy of the code used for the visualizations are provided for those interested in further exploring this new approach.