[tor-talk] Matryoshka: Are TOR holes intentional?
coderman at gmail.com
Thu Jun 18 20:45:31 UTC 2015
On 6/18/15, l.m <ter.one.leeboi at hush.com> wrote:
> All that padding means nothing if an adversary can introduce latency
> or gaps *at arbitrary* locations in a path. An adversary that can see
> your guard, and who can also see the guards traffic can introduce the
> gaps/latency in traffic at any point in your path. You may not even
> see the attack without being able to visualize end-to-end bandwidth
> statistics. It might be due to a routing problem at a particular node
> in the path. Solving this adversary isn't easy because they can hide
> behind the design of the internet.
this is where multi-path transports, which resist attacks against
traditional in-order or stream oriented transports - inherently
encumbered by serial datagram sequence. sadly, such transports also
require user space stacks and significant complexity. [see past
threads on multi-homed, multi-path SCTP w/ DTLS/IPsec telescopes, as
just one possible shape of could be...]
future ideas, plausible or no,
Matryoshka does not solve even the known problems, let alone beat Tor
at the privacy game.
... fun subjects for exploratory development, perhaps when Tor Project
gets a research endowment. :P
More information about the tor-talk