I doubt I'm your target market, so feel free to ignore me, but do you have a page that explains what you're really offering minus all the marketing hand waving? It sounds like it might be interesting but it's difficult to wade through the 10,000 ft view stuff.
For example, I just picked this mouseover that interested me:
The threat challenge page stops known threats and alerts infected humans that they need to take action.
Is there anything that elaborates on that? From a security perspective, I'm drawing a blank as to what a reverse proxy filter is achieving there. You're rewriting html destined for ddos zombies?
If a threat is detected (either because the IP has a bad reputation, or a request contains a malicious payload) then, depending on your security settings, instead of the request being passed to your web server it is answered by the proxy. The answer is a web page that, again depending on your security settings and the type of threat, includes a CAPTCHA. If the visitor passes the CAPTCHA then their session is marked as valid and they're allowed to pass through the proxy unhindered.
The nature of the system is that we're seeing data across tens of thousands of websites so we get smarter as we grow larger. We have a birdseye view into overall flow patterns and can spot attacks that are very difficult to see if you're only looking at your own logs from your own sites. For example, if the same IP hits multiple, unrelated CloudFlare websites then it is an indication that it is some kind of automated crawler. We can then look at whether it comes from a known, legitimate entity (e.g., Google) and also watch its behavior for other characteristics that indicate it may be a threat.
The chances are pretty low a scraper would hit multiple unrelated Cloudflare websites since Cloudflare is only used in very very few websites... Scrapers usually are interested in particular websites, they dun just scrape random sites.
What other characteristics can you detect? Can't really look at IP address, since ISPs such as AOL use the same IP address for the same user. Can't look at headers or referral strings since those can easily be faked. Also search engines such as Google have been known to use non-Google IP's to check if a site is cloaking or not. And you say you analyze the reputation of an IP - IP addresses for users change all the time. And many scrapers do use data farms/cloud services such as AWS, but a lot are moving to European data servers as well, and these IP addresses are harder to get reputation for (they're not in ARIN, etc).
For example, I just picked this mouseover that interested me:
The threat challenge page stops known threats and alerts infected humans that they need to take action.
Is there anything that elaborates on that? From a security perspective, I'm drawing a blank as to what a reverse proxy filter is achieving there. You're rewriting html destined for ddos zombies?