Information Security Stack Exchange is a question and answer site for information security professionals. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

A layman's question: If I don't err, with remailers or analogous systems one sends a message to a node of a chain of nodes, which collectively perform the entire transmission. However, the message to the first node of the chain must IMHO contain a valid IP-address of the sender. So, if the agencies have access to that very first seqment of the transmission, the sender could be found out (unless one sends from internetcafes or callshops). Is this correct?

share|improve this question
It sounds like you're asking this to confirm something else that you want to know. What is the purpose behind your question? Edit: As a basic answer: no, not when the source is actually an intermediary who is not disclosing the real source. – Luc Jul 4 '13 at 13:45
The purpose is to know whether use of remailers is "perfectly" safe and, if not, what would be the best to do. See also my comment to the post of qbi. – Mok-Kong Shen Jul 4 '13 at 14:23

Decent remailers who aim at providing anonymity use mix networks with encryption.

The sender wants to send his message M through the chain of remailers A, B and C (in that order). To do that, the sender encrypts the message M (complete with headers such as subject and destination address) with the public key of C. That encrypted message is then packed into another message with the address of C as destination address; that outer message is then encrypted again, this time with the public key of B. The result is again wrapped into a message with B's address as destination. That message is encrypted with A's public key, and then sent to A.

When A receives the message, it decrypts it and finds out an inner message to be sent to B. B itself will be able to decrypt it, and find the inner-inner message which is for C. C will decrypt that, and find the inner-inner-inner message M and the actual destination. That way, A won't be able to know where the message should go; also, C won't be able to know from where the message comes (because C received it from B, not from the sender). B knows nothing at all (from the point of view of B, the message is from A and sent to C).

This principle is also applied to HTTP requests in Tor, and can also be used to design electronic voting systems.

share|improve this answer

I'm assuming you're talking about SMTP mail here. If that's not the case, please edit your question and explain your context.

Since all nodes uses the same mechanism for forwarding the message, it is possible for the message originator to gorge a fake "received from" header in the SMTP envelop and thus suggest that the message actually originated form a different system.

Edit: To clarify after the OP comment, I want to add that this trick will not grant you anonymity, but if implemented correctly, it can provide you with plausible deniability.

share|improve this answer
Remailers don't always use SMTP. Mixminion for instance does also use TLS over TCP. – qbi Jul 4 '13 at 13:59
I suppose there is an misunderstanding. I mean: If I send a mail to anywhere from my PC, my IP-address is contained in the entire data package and that IP-address IMHO can't be faked. So if the transmission to the "first" node is tapped, the sender would be known. (I know that the normal sender address, which is in the form of a text string, could be altered, if one likes.) – Mok-Kong Shen Jul 4 '13 at 14:00
@qbi There is nothing in SMTP that mandates the use a of raw TCP connection. In fact, pretty much all SMTP servers out there are capable of using TLS for transport of SMTP messages. – Stephane Jul 4 '13 at 14:06

Based on your comment I updated my answer: You are talking about the transmission of your mail to the first node which is being tapped. The question is at what place the tapping device sits:

  • Your computer
    Well in this case you lost anyway. Because this probably means that the adversary has many more possibilities to look into your traffic. He would also find out the contents of your message in most cases.
  • A router between your computer and the receiver
    The router can look inside the IP packet. This packet has a destination address and a source address. So the adversary sees where the package comes from. If this tapping only happens at one or a few routers it might be helpful to enter a strict route. So that your packet don't go through the tapping devices.
  • receiving computer
    It is basically the same as with the route above. However there is a chance that on the receiving end only happens some tapping on higher protcols (SMTP etc.). In this case Stephane's answer applies. You could change the MAIL FROM in the envelope and are fine.

When the user also wants to hide his connection to the first remailer, an anonymisation tool like Tor or JonDonym might come in handy. However also in this case the adversary sees the connection to the first hop of Tor or JonDonym.

Another way would be to run an own Tor relay and use this relay for own use. So there are quite a lot of connections from that node to others. This gives the user plausible deniability.

share|improve this answer
If I don't err, it would thus seem to be safer to send messages from Internetcafes/callshops to such Usenet groups as alt.anonymous.message and let the recipient collect them from there according to certain distinguishing traits. – Mok-Kong Shen Jul 4 '13 at 14:11
Thanks. I am concerned about the location of the router of the first node that one sends the message to. Since such first nodes of remailer chains are certainly a very tiny fraction of all Internet nodes, it would be advantagesous IMHO for the agencies to concentrate their tapping activities there. – Mok-Kong Shen Jul 4 '13 at 14:35
[Addendum] The tapping points could be more conveniently situated at the providers of the first nodes of remailer chains, if I don't err. – Mok-Kong Shen Jul 4 '13 at 19:00

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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