Building an untrackable webserver

Question

I have a friend who is hosting some data one will be able to download from a server using two methods:

1. Using an app that they will first have to download onto their mobile device
2. Accessing the website directly

However, there's two important considerations for the data: it is very sensitive in that he doesn't want any "middle man" to be able to tell who got the data or where it came from (from server / ISP logs, etc...). It should also be accessible when visitors have extremely limited bandwidth around the world. The data should be publicly accessible, and easy to get by the people who know about its existence.

The ultimate goal here is to protect the visitor, so that middle men can't tell what content the visitor has requested, or from where he/she requested it. Basically, we want, as the title says, an untrackable webserver.

To help hide the web-site's identity from the middle man's logs, as well as to help protect the individual requesting the data, I'm thinking that one solution would be to build out several geographically-dispersed proxy servers that tie into a service like No-IP.com. I think it would also be good to somehow encrypt the data from point A to point B (something beyond simple SSL).

Coding considerations aside I'm wondering how you guys would go about achieving this type of security in the server & network side of the project.

Does my idea of using a service like no-ip.com, with a number of different proxy servers globally make any sense?

PS: I realize that some things in this post are vague. Unfortunately, I can't give any specifics as to what the data is, or why we need such a secure infrastructure - for the same exact reason that we need it at all. If you have concerns or questions, please feel free to reach out to me directly.

Answer 1

First, the no-ip idea is horrible. This doesn't hide you or your webservers in the slightest. They still actually have an IP and that IP is still tied to you in logs by your ISP. If you're afraid of anything showing up in logs, this isn't going to help as the same domain name is going to show up in the DNS server logs over and over again. This isn't a way to keep anonymous. It's simply meant as a way around needing static IP addresses.

Secondly, I don't know where you get the idea that SSL is not good enough. "Simple SSL" is mighty powerful in keeping data hidden. What makes you think that it's not good enough?

If you want to hide your servers somehow, it's just not going to happen. It's the structure of the internet. If you want your users to be able to hide what they're looking at, you should use a different method. Any method you deploy on your own will be traceable as that method is only good for accessing your site. Have your users use Tor or build an app that uses it on the back end. It's probably one of the few easy, reliable ways to accomplish what you want.

Answer 2

I would get them to VPN into a point that you control, and then proxy SSL connections over the VPN.

The encrypted VPN hides any access logs from anyone in the middle, and then SSL protects the actual data being transferred from anyone who might be inside that VPN tunnel (i.e. the VPN host itself).

You could even use an SSL VPN to simplify setup, but SSL VPNs are usually tied to specific platforms (Windows, usually)

There's not really any such thing as an un-trackable web server. Hollywood has lied to you in that regards. "OMFG the web server's IP address changes every 10 seconds, so there's no way we can block the transmission!" is utter rubbish.

Answer 3

Hosting an .onion service is what solves your problem. Tor is the only solution to this problem. Data on Tor network is encrypted and is decrypted only at relay nodes so middleman gets nothing, no trace. Frequently changing Tor identity will help great but the service address will be needed to communicated every time new identity is issued.