My question is about CDNs. As I understand it, data flow looks something like this:
browser ==(https)==> CDN ==(https)==> originServer
In theory, this configuration should provide quicker retrieval time for the static .html, .js, .css, and .png files that comprise my app. All of these items are static and non-confidential, by the way. From the user's perspective, they live at https://www.myDomain.com.
But there's also dynamic data that's exchanged via https://ajax-www.myDomain.com. That content actually is confidential and I'd like to protect it from snooping eyes. It occurs to me that any employee working at the CDN could read the confidential 'ajax-www' content if he so desired, seeing as how the whole point of the CDN (as I understand it) is to act as a HTTPS termination point, cache data when possible, and pass it on.
Of course, I could circumvent the CDN entirely for 'ajax-www' calls. After all, the data in question is dynamic and can't be cached anyway. But then I'd be revealing the location of the origin server to the outside world and more importantly, I'd be losing the DDOS protection that the CDN provides.
The discussion here recommends that CDNs should be avoided for session-specific data but is there a viable alternative?
Is this even something worth worrying about? Has there ever been a (known) case of a CDN snooping on client traffic, either of their own volition or as a result of government pressure? Casual searching doesn't turn up anything. CDNs seem to be considered beyond reproach. But I just figured I'd ask -- the specific CDN that's under consideration is based out of Dubai so it's not like we'd have any recourse if snooping were to occur.
Sorry for the basic nature of this question. I'm kind of new to this stuff. Looking forward to hearing your thoughts.