Re: [tor-talk] ATOMIC BANJO and LEVITATION used by CSE

[Paul Syverson <paul.syverson@xxxxxxxxxxxx>]

On Mon, Feb 02, 2015 at 01:14:15PM -0800, spencerone@xxxxxxxxxxxxxxx wrote:
> >mirimir[at]riseup.net:
> >>spencerone[at]openmailbox.org:
> >>>paul.syverson[at]nrl.navy.mil:
> >>>>See p. 129 of
> >>>>http://www.acsac.org/2011/program/keynotes/syverson.pdf
> >>>>also
> >>>>https://www.torproject.org/docs/faq.html.en#WhyCalledTor
> >>>>
> >>>>aloha,
> >>>>Paul
> >>>>
> >>>>(Note, the German meaning of 'Tor' mentioned in the FAQ is discussed
> >>>>in the "A Peel of Onion"
> >>>>paper, the Turkish meaning is apparently a fine-meshed net.)
> >>>
> >>>Awesome, that kinda makes sense, tough, given that Tor is THE onion
> >>>router, I think referring to Tor as TOR is still accurate :)
> >
> >No, Tor is not "THE onion router". It's _an_ onion router :)
> >
> >You should rather say Tor is not _the_ onion routing, it's _an_ onion
> >routing: cf. p. 129 again---except this is not the 'the' of definite
> >description but the 'the' of "The original and still the best". (I got
> >this phrase from Roger Needham in 1993.  He was talking about BAN
> >logic, and said he got the phrase from a shoe polish tin.)
> >
> >aloha,
> >Paul
> 
> "The original and still the best" is what THE means to me :) 

Great. So we were on the same page already.

> Though [ing] vs [er] seems debatable since Tor is a thing that does
> onion-like layered routing.

Hopefully the paper I just sent you helped. Basically what we called
'onion' in pre-Tor versions of onion routing from NRL doesn't exist in
Tor. An onion was just layers, no middle. It was used to build the
circuit. What came after the circuit was built had content in the
middle of the layers, so was not truly an onion.  (Also the onion's
layers were public-key crypto, what came later was symmetric-key.)  We
changed how circuits were built in the Tor design, but kept the name
from the earlier designs since the core concept was the same.  Most
importantly wrt the name: an onion router is just a Tor relay.  Tor's
not just an individual relay. It's the whole system: all the relays,
the directory system, the client software, etc.

> 
> But, to understand more, are the other onion routing projects implementing
> their own onion routing protocol or are they implementing Tor? I could
> investigate this myself but I don't know enough to figure out the difference
> unless explicitly stated.

The point was that there was a bunch of stuff we started doing at NRL
in 1995 we called "onion routing" including what we eventually called
Tor.  Some people not at NRL designed, and in some cases built, other
systems using the same onion routing principles (e.g. the Freedom
Network that Zero Knowledge Systems ran c. 2000-2001, Iron Key had its
own private onion routing network, plus lots of academic paper
designs, etc.)  The contrast in the name, why it was _the_ onion
routing, was comparing Tor to other non-NRL pre-Tor onion routing
systems. Tor was part of the original set of projects and design goals
at NRL (which had a bunch of iterations and revisions so that none
after early 1996 was the original original onion routing). None of the
other non-NRL systems from the pre-Tor days would have been implementing
Tor. Although Tor imported a key idea from at least Cebolla and
probably some other ideas from others as well (or learned things _not_
to do from some of them).

HTH,
Paul
-- 
tor-talk mailing list - tor-talk@xxxxxxxxxxxxxxxxxxxx
To unsubscribe or change other settings go to
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk