Transferring data quickly via Tor

[grarpamp]

On Wed, Jan 25, 2017 at 2:43 PM, John Newman <jnn at synfin.org> wrote:
>
>> On Jan 24, 2017, at 3:52 AM, Mirimir <mirimir at riseup.net> wrote:
>>
>> What I need is help understanding the privacy implications. I'm going to
>> explore possibilities for moving long transfers randomly across sets of
>> subflows. Using "roundrobin" as mptcp_scheduler instead of "fullmesh"
>> would be a start. That would also spread load across more relays.

>> Another possibility is aggregating OnionCat and GarliCat links, so
>> transfers would be split between Tor and I2P.

This sickness among us :)

> At the moment MPTCP is not implemented in FreeBSD and i don't have a linux machine particularly convenient to play with this on, although that should change soon.
>
> I talked to some people on #freebsd and there is an MPTCP source tree, but it's essentially a fork at the moment, total PITA to get merged into a running system, probably 11.x only (im still at 10.3-release on my handful of machines).

It's coming along, people can compile and play with it in VM...

https://bitbucket.org/nw-swin/caia-mptcp-freebsd
http://caia.swin.edu.au/newtcp/mptcp/
http://caia.swin.edu.au/newtcp/
https://www.freebsdfoundation.org/project/multipath-tcp-for-freebsd/

http://blog.multipath-tcp.org/
http://multipath-tcp.org/
http://www.theregister.co.uk/2014/08/07/facebook_wants_linux_networking_as_good_as_freebsd/

# Use Cases and Operational Experience with Multipath TCP
https://tools.ietf.org/html/rfc8041
https://tools.ietf.org/wg/mptcp/
http://www.theregister.co.uk/2014/07/31/multipath_tcp_will_bork_your_network_probes_flummox_your_firewalls/