The dark web is like a lair for hackers/criminals/etc. But has anyone, like the FBI or CIA ever found the main server of the dark web? Where is the main server located? I thought all Internet data flows are being watched by many government institutes like the FBI/CIA. Why have they not closed down the main server of the dark web itself?

I mean the dark web is actually a part of the "healthy"/"surface" web, right? Why did they not get caught by the main server provider?

  • 2
    Why do you think there is a single server? Why would it be a single server? – schroeder Dec 26 '17 at 16:47
  • When you are talking about the TOR network (which is not what IT professionals call "dark web" but what many laypeople think of when they do), then there is no "main server". – Philipp Dec 26 '17 at 17:31

The dark web is like a lair for hackers/criminals/etc.

That's an oversimplification, but we'll go with it.

But has anyone, like FBI or CIA ever found the main server of the dark web? Where is the main server located? I thought all Internet data flows are being watched by many government institutes like FBI/CIA. Why have they not chased the main server of the dark web itself?

I mean the dark web is actually a part of the "healthy"/"surface" web right ? why they did not get caught by the main server provider ?

We apparently need to cover a broad number of basics here.

First, there is no main server of the public web. That is not how the web works. I can't go into that much detail in an SE answer, but your computer generally uses (the distributed) DNS to turn a human-readable hostname into an IP address, and then that gets routed across physical wires based on (again, distributed) routing protocols and data until it gets to the server it's supposed to get to.

The so-called dark web is structured similarly - different people run different web servers, and traffic has to go to the right places. The difference is in how you get there.

There's an anonymizing technology called Tor. When used to connect to the public web, a tor user's traffic goes through three different intermediate servers before trying to get to the requested website. The first node knows who the user is, but not what they were requesting (it's encrypted). The third node knows what's being requested, but not who requested it. And the middle node knows essentially nothing. This process prevents the website and other MitM snoopers like the NSA from knowing who is requesting that website.

(Insert a bunch of caveats here around anonymity with tor, attacks on the protocol, safe browsing, and general advice to please go look up more than this comment before relying on it.)

Tor also provides a feature known as hidden services, where the website operator also sets up a circuit of three nodes to connect to. A tor user's circuit meets up with the website's circuit, and traffic flows through all six of the nodes without anyone knowing who the other parties are. This is how the deep web works.

The only question then is how the two meet up in the first place, and that uses yet another distributed piece of technology, a distributed hash table built into tor. So there's no centralized server to shut down here, either.

The weak point here is that the website operator needs to not leak their identity or the server's location on accident, which can happen through all sorts of misconfiguration or human error. But even if they do, that is only that one site that gets shut down, not some giant shutdown of the network. Also, anonymity is not illegal, so there is no legal reason for the feds to shut down tor overall, even if they could. It in fact started as a US government project to assist people in oppressive regimes.


This question is interesting in that it reveals some assumptions people have about the internet as whole.

First of all, the basic premise of this question is, for lack of a better term, wrong. The internet as a whole was designed by the US Department of Defense to be resilient to a Soviet nuclear attack and to withstand elimination of nodes without losing communication by bypassing missing nodes. As such, there isn't a lot of central control of the internet because that would eliminate its main purpose (resiliency in the event of a nuclear strike).

Secondly, the level of traffic on the internet is HUGE. Noting Arminius's comment above, the FBI/CIA/what have you would need to sift through an amazing amount of traffic on the web to get to the deep web traffic. Cisco estimates that in 2017 this was 96,054 petabytes per month.
https://en.wikipedia.org/wiki/Internet_traffic It would take a lot of computing resources to comb through that, to put it mildly.

Even if there were a main deep web server, as the question asks, finding it would be a challenging. IP addresses are logical addresses and don't reveal much of anything. You can track ownership of the address through ICANN, but there doesn't need to be real person responsible for it and they can be located anywhere in the world, processed through layers of shell companies. You would be able to track some info through the ISP, but that assumes that other things (microwave relay, running a wire) haven't been done to mask the location of the server in question. And, were the alleged central server to be located in a country where law enforcement is lax, corrupt, or non-existent, there's almost nothing that could be done. Sure, we'd know where it is, but we won't get any cooperation from law enforcement there.

So, to answer your question:

  1. The deep web is part of the internet as a whole; it's not a separate internet
  2. There is no central server for the internet, much less for the deep web
  3. The amount of traffic on the internet is huge, making it logistically impossible to filter through it all and find something if you don't know where to start
  4. Even if such a server did exist, it would be difficult to find and if out of the reach of US law enforcement, impossible to seize.
  • 1
    But your scenario for finding the server depends on the idea that you don't know what you are looking for. If you can connect to the server, it becomes a much more targetted task to physically locate it (with law enforcement and ISP cooperation). Searching for a server that you do not know is the herculean task. – schroeder Dec 26 '17 at 15:21
  • @schroeder good point, noting one assumption I made that may be in error. – baldPrussian Dec 26 '17 at 15:23
  • This answer doesn't address caveats intrinsic to deanonymizing "dark web" services, such as Tor hidden services, I2P eepsites, or Freenet freesites. The reason it is hard to find such a server is not because the internet is huge, but because these services use anonymizing techniques to efficiently hide their location while still being addressable with the right technology. – forest Dec 27 '17 at 2:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.