Rank lists of domains, IPs and other criteria are widely used for security and internet applications. Yet research shows that with the publicly available rank lists widely in use today, such as Amazon’s Alexa Top 1 Million domains, rankings can vary considerably over just a few days. In particular, ranks based on observation counts in a network can be affected by incorrectly collected data, congestion in the network, seasonality, user trends, and other factors. This variability can have negative real-world effects on system security and performance.
For these reasons, Infoblox has implemented a technique that relies on rank lists that allows us to avoid selection bias.
- Our ranking system for domains provides not only a range of plausible ranks but also the most likely single rank observed during a particular period.
- We use statistical inference to create a statistically significant rank list that describes a domain’s stability and true rank over time.
- We combine data collected over a period of time and use that data to define a rank’s confidence interval for each domain.
Complete the form to receive your copy of the new White Paper from Infoblox detailing the methodology underlying InfoRanks – The Infoblox Ranking Service.