Re: <nettime> FW: When technology is utilized against us.

[Evan Buswell <ebuswell@gmail.com>]

Nick points out stuff I was skimming over in my analysis.  That's
good, since I left one side out.  See below for responses, however.

On Wed, Jun 24, 2009 at 2:42 PM, Nick <nick@njw.me.uk> wrote:

> Hi list. I'm new here; recently found this place from a reference to
> a Jeffrey Juris interview, and I'm very interested in the
> discussionn here.
>
> I am also in agreement with Morlock's analysis, as harsh and
> condescending as the tone may have been.
>
> I have some disagreements with Evan's points though:
>
> Quoth Evan Buswell:
> > On the internet, the main problem is not actually encryption,
> > silencing, it is trust. We can't see each other; we have little
> > tangible physical relationship. Nobody can be sure whether their
> > messages are consistently going to the same place or not. If that
> > problem is not solved, then encryption is just ensuring that no other
> > government agents are intercepting the communication between you and
> > the government agent that, unbeknown to you, you are speaking to,
> > and is in turn speaking to somebody else on your behalf. This is the
> > certificate hell that all internet security has entered into. For the
> > company I worked for, that issue boiled down to everyone trusting the
> > service providing company to sort out the identity of everybody else.
> > And of course, the more completely the problem is solved, the more
> > complete control over everything that one company has, with the limit
> > being just about the same hypothetical vulnerability that everyone
> > else has sending it all in the clear and trusting the ISPs forwarding
> > the packets not to legally or illegally be monitoring packet flow.
>
> There are a host of schemes to prevent Man In The Middle attacks;
> see OpenPGP signing or ZRTP for example.
>
> And the trust issue is been addressed by 'web of trust' models, as
> practised with OpenPGP, or Freenet. No centralisation needed. At
> all. Â(though voluntarily some centralised services can be used, for
> convenience, e.g. OpenPGP keyservers).

MITM attacks are only prevented by proper trust based on certificates.
That was my point, though I see now how that may not have been
underlined in my (probably too negative) text.

OpenPGP/various "web of trust" models are all great stuff to bring up,
as the centralized model is certainly not the only one.  But one only
has to look at such things as "key signing parties" to realize that
this is a social solution, not a purely technological one.
Definitely, though, my larger point is not that there's nothing we can
do, but that our lack of a widespread solution is for reasons other
than laziness.  Establishing trust is a social phenomenon.  We have no
agreed-upon encoded representation of that phenomenon because the
interface between our internet and non-internet practice of trust and
identity is complicated.

Freenet gives semi-anonymity with semi-trust.  You know that a key
continues to belong to who it first belonged to, but not much else.
If you crack one message source, you know all the messages which that
author published and the person effectively loses their "anonymous"
identity.  Freenet's (and other's) idea of anonymous-but-establishable
identity is really interesting, but I hope it is apparent the way that
idea radically differs from our everyday ideas about what it means to
be anonymous or identifiable.

Again, I don't want to imply by any of this either that there are not
already solutions one could use or that better solutions cannot be
found, just that there are much more complicated things involved.  I
think some Freenet-like model is fantastic, but right now it's only a
special solution (for both social and technical reasons), and as such
not available in public places like internet cafes, from which a lot
of activists end up posting their messages.  Note also that the
practice of posting in a cafe in most cases solves the problem of
anonymity better than technology and without any sort of technical
solution backing it.

> > Which brings me to the second point: anonymity is a very different
> > problem than encryption. The messages being posted from Iran are
> > public, by design. Because all IP records both source and destination
> > address, in the absence of random message delays, all that the Iranian
> > government has to do is monitor packet flow and correlate that with
> > the times when suspicious messages publicly appear. It doesn't
> > actually have to read anything that those packets contain. I have no
> > idea what "deep packet analysis" is supposed to mean, but I would
> > imagine that the analysis I'm talking about here is all Iran is doing.
> > It seems like probably all they need to do.
>
> Again, isn't this solved by using web of trust, associated to an
> identity which is different to one which could readily incriminate
> 'the real' you?

Not really.  That part only establishes a fictional identity, it
doesn't de-link it from your real one.  Encryption isn't really
required since these are not secret messages.  Freenet (for example),
I think does this by things like packet delay and random peer-to-peer
forwarding and all sorts of other stuff that works really hard to
protect the user's identity, but also vastly degrades the performance
of the network.  You can forget about real-time updates and
broadcasts---though the delay doesn't have to be huge.  Plus, any
packet-forwarding anonymizer's strength (and usually performance) is
always going to be proportional to the number of people using it who
can't be eliminated as suspects, which is again, a social problem that
isn't so much about the laziness of a subversive person as it is about
the apathy of those not engaged in any subversive messaging
themselves.

> And deep packet inspection/analysis is looking inside an IP packet
> at the content, e.g. if a series of HTTP packets contained a video,
> it could be fingerprinted and compared against a database of
> 'forbidden' content, and the stream could be closed by a hostile
> ISP.

Gotcha.  Then this is more about blocking information than identifying
its source?  Yeah, any cipher should bypass most of that (including
https), even with government-known keys, just due to the extra
computational load.

I guess overall, I do appreciate a solutions-oriented rather than an
absolutely-shocked response to all this; that point is well-taken.
But we can only really get on the path to creating solutions when we
acknowledge the complexity of how all of this works.  We're talking
about trust and identity, stuff like the scarlet pimpernel or batman
with masks and dual identities and sabotaged expectations and creating
secret spaces for whispering and riding around in random directions on
bicycles with morse code transmitters wired to pre-taped codes, not
just uncomfortably twiddling with bits.

> I'll be interested to hear what others on the list have to say.
>
> Nick White
>
> --
> GPG Â Â : 0x04E4653F / 9732 D7C7 A441 D79E FDF0 94F6 1F48 5674 04E4 653F
> IM /OTR : njw@jabber.org / 7E3C82CC D6AB2CEA E8000300 E429A122 D984111B
> SIP/ZRTP: njw@ekiga.net
> WWW Â Â : http://www.njw.me.uk


#  distributed via <nettime>: no commercial use without permission
#  <nettime>  is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mail.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nettime@kein.org