I don’t think it’s arguable that nginx is sometimes used as a reverse proxy (we use it as a load balancer in front of Java application servers at work, for instance), though I doubt anyone knows how often or what it’s proxying. (It might be possible to fingerprint the underlying server, which could give interesting results.) The author correctly observes that Server: headers don’t tell the full story, but as you note that observation is necessary but not sufficient for their conclusion.
[Citation Needed]
I don’t think it’s arguable that nginx is sometimes used as a reverse proxy (we use it as a load balancer in front of Java application servers at work, for instance), though I doubt anyone knows how often or what it’s proxying. (It might be possible to fingerprint the underlying server, which could give interesting results.) The author correctly observes that
Server:headers don’t tell the full story, but as you note that observation is necessary but not sufficient for their conclusion.Thanks for this article - it’s also interesting to compare it to NetCraft Server Survey who have been data mining the internet since 1995 :~)
Especially when you compare it to the 2003 survey