[jarusl at cs.northwestern.edu: Tor proxy for Virtual Machines]

[Kyle Williams]

I've seen this and it is very interesting.  I like how reference 16 
points to JanusVM, which basically does the same thing.  JanusVM also 
works with Qemu that has been patched to work with LibPCAP.

However, the crutch with both projects being "... that you have the PCAP 
and LIBNET libraries installed."  Which really isn't much.

With JanusVM the host OS is using the guest OS to transport traffic over 
Tor.
With TorVTL the guest OS is using Tor from the host OS to transport 
traffic.  I would imagine you could run this inside a VM too if one so 
desired.

Very nice. :)

- Kyle


Nick Mathewson wrote:
> Forwarding this with permission.  It looks like interesting work,
> especially for people pursuing VM-based anonymization strategies.
> 
> ----- Forwarded message from "John R. Lange" <jarusl at cs.northwestern.edu> -----
> 
> From: "John R. Lange" <jarusl at cs.northwestern.edu>
> To: tor-volunteer at torproject.org
> Subject: Tor proxy for Virtual Machines
> X-Spam-Level: 
> 
> Hi,
> 
> As part of one of our research projects, I put together a small proxy tool 
> that anonymizes all TCP and DNS traffic originating in a virtual machine 
> (at least with VMWare or Xen). It currently runs under linux, but contains 
> very early support for Windows.
> 
> It plugs in beneath the Virtual Machine Monitor, so it is capable of 
> providing TOR support for any application+OS combination without any 
> configuration or special proxy tool needing to be installed in the guest 
> environment.
> 
> It works by configuring a VM's network adapter to connect to a host-only 
> network, where none of the packets are ever transmitted out of the host 
> machine. Every ethernet packet coming from the VM is then captured and 
> translated into SOCKs traffic that is forwarded to a TOR proxy. If the 
> packet is not supported then it is simply dropped and is never transmitted 
> on the network.
> 
> More info on the tool can be found here:
> http://www.artifex.org/~jarusl/TorVTL/
> 
> While the paper can be found here:
> http://www.artifex.org/~jarusl/research/pubs/hpdc07-vtl.pdf
> 
> I'm unsure whether this will be of any interest to people, but I figured 
> I'd let you know.
> 
> regards
> --Jack
> Jack Lange ; NU CS ; jarusl at cs.northwestern.edu
> http://www.cs.northwestern.edu
> 
> ----- End forwarded message -----
>