Super User is a question and answer site for computer enthusiasts and power users. 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

Does it make any sense it terms of perfomance and latency?As far as i know TOR is a transports layer protocol trying to encapsulate the cells with onion routing in order to enhanse privacy.Since sip is widely used as a session protocol for voip calls what do you believe about perfomances of that incarnation?

share|improve this question

Remember that just using TOR for SIP (and not for actual data) isn't good for anonymizing your connection.

Using voip over TOR is definitely foolish, because TOR is slow by it's nature. Your traffic will go through multiple nodes, some being normal home broadband connections. Each connection adds latency, in my experience with rather much (latencies over 200ms are more than common, and it's not my end).

Performance (bandwidth) really depends on your luck. You may get fast connection, or may not. And it also changes regularly, so fast connection may change to horribly slow. Remember that your connection is as slow as slowest link between you and exit node.

share|improve this answer
What else is beeing used to provide anonymity to voip traffic? – curious Mar 1 '11 at 19:10

I made my thesis this year, whose subject was SIP over Tor. Soon you will be able to read a paper we prepare about this.

To the point, I made measurements of delay when using SIP over Tor and I could say the results where quite interesting. Both caller's and callee's servers should use Tor, so if they have a really fast internet connection, then delay is not as much as you think.

The real problem of SIP is that in a call establishment, many untrusted proxies, may intervene. This is where you need a privacy protection system, like Tor.

Let's suppose, Bob wants to call Alice. His smartphone is using Orbot to connect to the Tor Network. Alice is registered to a Server A. If Server A uses Tor for his connections, then he will retransmit the INVITE message to Alice, through Tor. After that, Alice will response to the last Tor node, who sent the request. In this case, both sides are protected, while only the caller and callee's Server use Tor.

As everyone knows, Tor works only on TCP while VoIP services work on UDP and that offers a slightly longer delay. But this delay in real action is NOT noticeable. In addition, Tor will be used only on Call establishment. After the call is established, users can talk to each other without using Tor. It's not that reasonable to use Tor in a direct "user to user" connection.

share|improve this answer
i would like to see your report – curious Feb 9 '14 at 17:41
@Jack What is the point of using Tor for establishment but not communication? What use-case would it be useful to have only anonymous call establishment but you loose anonymity once you begin communicating? – Scott Chamberlain Feb 9 '14 at 17:55
@ScottChamberlain Well, the implementation of SIP over Tor exists and works nice, this was my answer to the exact question.I guess someone could use Tor even for communicating, but this would not lead to efficient calls between the users (consider a group call with more than 2 users). Our security concern was that during the call's establishment many untrusted proxies will intervene and can obviously see the content of SIP packets, while we don't want them to. – Jack Feb 10 '14 at 14:17
I see, the threat model was to protect a unsecured handshake, not prevent backtracking of the caller, that makes sense. – Scott Chamberlain Feb 10 '14 at 14:31
@curious I promised you a report. Here is the article, which I will present to ARES conference this september. – Jack Jul 22 '14 at 14:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .