tahoe-lafs/docs/tor.rst

Using Tahoe-LAFS with Tor

1. Use cases
2. Native Tor integration for Tahoe-LAFS
3. Software Dependencies
4. Configuration
5. Performance and security issues of Tor Hidden Services
6. Torsocks: the old way of configuring Tahoe-LAFS to use Tor

Use cases

Tor is an anonymizing network used to help hide the identity of internet clients and servers. Please see the Tor Project's website for more information: https://www.torproject.org/

There are three potential use-cases for Tahoe-LAFS on the client side:

1. User does not care to protect their anonymity or to connect to anonymous storage servers. This document is not useful to you... so stop reading.
2. User does not care to protect their anonymity but they wish to connect to Tahoe-LAFS storage servers which are accessbile only via Tor Hidden Services.
3. User wishes to always use Tor to protect their anonymity when connecting to Tahoe-LAFS storage grids (whether or not the storage servers are Tor Hidden Services) [*].

For Tahoe-LAFS storage servers there are three use-cases:

1. Storage server operator does not care to protect their own anonymity nor to help the clients protect theirs. Stop reading this document and run your Tahoe-LAFS storage server using publicly routed TCP/IP.

2. The operator does not require anonymity for his storage server, but he wants it to be available over both publicly routed TCP/IP and through Tor Hidden Services. One possible reason to do this is because being reachable through Tor Hidden Services is a convenient way to bypass NAT or firewall that prevents publicly routed TCP/IP connections to your server. Another is that making your storage server reachable through Tor Hidden Services can provide better protection for your clients who themselves use Tor to protect their anonymity [*].

   See this Tor Project page for more information about Tor Hidden Services: https://www.torproject.org/docs/hidden-services.html.en

3. The operator wishes to protect their anonymity by making their Tahoe server accessible only via Tor Hidden Services.

Native Tor integration for Tahoe-LAFS

Native Tor integration for Tahoe-LAFS utilizes the Twisted endpoints API:: * https://twistedmatrix.com/documents/current/core/howto/endpoints.html

Twisted's endpoint parser plugin system is extensible via installing additional Twisted packages. The native Tor integration for Tahoe-LAFS uses endpoint and parser plugins from the txsocksx and txtorcon modules. Although the Twisted endpoint API is very flexible it is missing a feature so that servers can be written in an endpoint agnostic style. We've opened a Twisted trac ticket for this feature here:: * https://twistedmatrix.com/trac/ticket/7603

Once this ticket is resolved then an additional changes can be made to Foolscap so that it's server side API is completely endpoint agnostic which will allow users to easily to use Tahoe-LAFS with many protocols on the server side.

txsocksx will try to use the system tor's SOCKS port if available; attempts are made on ports 9050 and 9151. Currently the maintainer of txsocksx has not merged in our code for the Tor client endpoint. We'll use this branch until the Tor endpoint code is merged upstream:: * https://github.com/david415/txsocksx/tree/endpoint_parsers_retry_socks

txtorcon will use the system tor control port to configure Tor Hidden Services pending resolution of tor trac ticket 11291:: * https://trac.torproject.org/projects/tor/ticket/11291

See also Tahoe-LAFS Tor related tickets #1010 and #517.

Software Dependencies

• Tor (tor) must be installed. See here: https://www.torproject.org/docs/installguide.html.en

• The "Tor-friendly" branch of txsocksx must be installed ( Once this is merged then you can use upstream txsocksx; https://github.com/habnabit/txsocksx/pull/8 ) :

  pip install git+https://github.com/david415/txsocksx.git

• txtorcon must be installed :

  pip install txtorcon

Once these software dependencies are installed and the Tahoe-LAFS node is restarted, then no further configuration is necessary for "unsafe" Tor connectivity to other Tahoe-LAFS nodes (client use-case 2 from Use cases, above).

In order to implement client use-case 3 or server use-cases 2 or 3, further configuration is necessary.

Configuration

[node] anonymize = (boolean, optional)

This specifies two changes in behavior:

1. Transform all non-Tor client endpoints into Tor client endpoints.
2. Force tub.location to be set to "safe" values.

This option is critical to preserving the client's anonymity (client use-case 3 from Use cases, above). It is also necessary to preserve a server's anonymity (server use-case 3).

When anonymize is set to true then tub.location does not need to be specified... and it is an error to specify a tub.location value that contains anything other than "UNREACHABLE" or a Tor Hidden Service Twisted endpoint descriptor string.

If server use-case 2 from Use cases above is desired then you can set tub.location to a Tor Hidden Service endpoint string AND "AUTODETECT" like this:: tub.location = "AUTODETECT,onion:80:hiddenServiceDir=/var/lib/tor/my_service"

It is an error to specify a tub.location value that contains "AUTODETECT" when anonymize is also set to true.

Operators of Tahoe-LAFS storage servers wishing to protect the identity of their storage server should set anonymize to true and specify a Tor Hidden Service endpoint descriptor string for the tub.location value in the tahoe.cfg like this:: tub.location = "onion:80:hiddenServiceDir=/var/lib/tor/my_service"

Setting this configuration option is necessary for Server use-cases 2 and 3 (from Use cases, above).

Performance and security issues of Tor Hidden Services

If you are running a server which does not itself need to be anonymous, should you make it reachable as a Tor Hidden Service or not? Or should you make it reachable both as a Tor Hidden Service and as a publicly traceable TCP/IP server?

There are several trade-offs effected by this decision.

NAT/Firewall penetration

Making a server be reachable as a Tor Hidden Service makes it reachable even if there are NATs or firewalls preventing direct TCP/IP connections to the server.

Anonymity

Making a Tahoe-LAFS server accessible only via Tor Hidden Services can be used to guarantee that the Tahoe-LAFS clients use Tor to connect. This prevents misconfigured clients from accidentally de-anonymizing themselves by connecting to your server through the traceable Internet.

Also, interaction, through Tor, with a Tor Hidden Service may be more protected from network traffic analysis than interaction, through Tor, with a publicly traceable TCP/IP server.

XXX is there a document maintained by Tor developers which substantiates or refutes this belief? If so we need to link to it. If not, then maybe we should explain more here why we think this?

Performance

A client connecting to a Tahoe-LAFS server through Tor incurs substantially higher latency and sometimes worse throughput than the same client connecting to the same server over a normal traceable TCP/IP connection.

A client connecting to a Tahoe-LAFS server which is a Tor Hidden Service incurs much more latency and probably worse throughput.

Positive and negative effects on other Tor users

Sending your Tahoe-LAFS traffic over Tor adds cover traffic for other Tor users who are also transmitting bulk data. So that is good for them -- increasing their anonymity.

However, it makes the performance of other Tor users' interactive sessions -- e.g. ssh sessions -- much worse. This is because Tor doesn't currently have any prioritization or quality-of-service features, so someone else's ssh keystrokes may have to wait in line while your bulk file contents get transmitted. The added delay might make other people's interactive sessions unusable.

Both of these effects are doubled if you upload or download files to a Tor Hidden Service, as compared to if you upload or download files over Tor to a publicly traceable TCP/IP server.

Torsocks: the old way of configuring Tahoe-LAFS to use Tor

Before the native Tor integration for Tahoe-LAFS, users would use Torsocks. Please see these pages for more information about Torsocks:: * https://code.google.com/p/torsocks/ * https://trac.torproject.org/projects/tor/wiki/doc/torsocks * https://github.com/dgoulet/torsocks/

Starting And Stopping

Assuming you have your Tahoe-LAFS node directory placed in ~/.tahoe, use Torsocks to start Tahoe like this:: usewithtor tahoe start

Likewise if restarting, then with Torsocks like this::

usewithtor tahoe restart

After Tahoe is started, additional Tahoe commandline commands will not need to be executed with Torsocks because the Tahoe gateway long running process handles all the network connectivity.

Configuration

Before Tahoe-LAFS had native Tor integration it would deanonymize the user if a tub.location value is not set. This is because Tahoe-LAFS at that time defaulted to autodetecting the external IP interface and announced that IP address to the server.

Tahoe-LAFS + Torsocks client configuration:

* Run a node using ``torsocks``, in client-only mode (i.e. we can

make outbound connections, but other nodes will not be able to connect to us). The literal 'client.fakelocation' will not resolve, but will serve as a reminder to human observers that this node cannot be reached. "Don't call us.. we'll call you":

tub.port = 8098
tub.location = client.fakelocation:0

Tahoe-LAFS + Torsocks storage server configuration:

* Run a node behind a Tor proxy, and make the server available as a Tor

"hidden service". (This assumes that other clients are running their node with torsocks, such that they are prepared to connect to a .onion address.) The hidden service must first be configured in Tor, by giving it a local port number and then obtaining a .onion name, using something in the torrc file like:

HiddenServiceDir /var/lib/tor/hidden_services/tahoe
HiddenServicePort 29212 127.0.0.1:8098

once Tor is restarted, the .onion hostname will be in /var/lib/tor/hidden_services/tahoe/hostname. Then set up your tahoe.cfg like:

tub.port = 8098
tub.location = ualhejtq2p7ohfbb.onion:29212

Troubleshooting

On some NetBSD systems, torsocks may segfault:

$ torsocks telnet www.google.com 80
Segmentation fault (core dumped)

and backtraces show looping libc and syscalls:

#7198 0xbbbda26e in *__socket30 (domain=2, type=1, protocol=6) at socket.c:64
#7199 0xbb84baf9 in socket () from /usr/lib/libc.so.12
#7200 0xbbbda19b in tsocks_socket (domain=2, type=1, protocol=6) at socket.c:56
#7201 0xbbbda26e in *__socket30 (domain=2, type=1, protocol=6) at socket.c:64
#7202 0xbb84baf9 in socket () from /usr/lib/libc.so.12
[...etc...]

This has to do with the nature of the torsocks socket() call wrapper being unaware of NetBSD's internal binary backwards compatibility.

Information on a the first parts of a solution patch can be found in a tor-dev thread here from Thomas Klausner:

• https://lists.torproject.org/pipermail/tor-dev/2013-November/005741.html

As of this writing, torsocks still exists in the pkgsrc wip tree here:

• http://pkgsrc.se/wip/torsocks

but the NetBSD-specific patches have been merged upstream into torsocks as of commitid 6adfba809267d9c217906d6974468db22293ab9b:

• https://gitweb.torproject.org/torsocks.git/commit/6adfba809267d9c217906d6974468db22293ab9b