ietf
[Top] [All Lists]

Re: [Fwd: I-D Action: draft-carpenter-prismatic-reflections-00.txt]

2013-09-22 16:08:23


--On Sunday, 22 September, 2013 12:59 -0400 Paul Wouters
<paul(_at_)cypherpunks(_dot_)ca> wrote:

Except that essentially all services other than email have
gained popularity  in centralized form, including IM.

Note that decentralising makes you less anonymous. If everyone
runs
their own jabber service with TLS and OTR, you are less
anonymous than
today. So "decentralising" is not a solution on its own for
meta-data
tracking.

Perhaps more generally, there may be tradeoffs between content
privacy and tracking who is talking with whom.  For the former,
decentralization is valuable because efforts to compromise the
endpoints and messages stored on them without leaving tracks is
harder.  In particular, if I run some node in a highly
distributed environment, a court order demanding content or logs
(or a call "asking" that I "cooperate") in disclosing data,
keys, etc., would be very difficult to keep secret from me (even
if it prevented me from telling my friends/ peers).   And a lot
more of those court orders or note would be required than in a
centralized environment.  On the other hand, as you point out,
traffic monitoring is lots easier if IP addresses identify
people or even small clusters of people.

The other interesting aspect of the problem is that, if we want
to get serious about distributing applications down to very
small scale, part of that effort is, I believe necessarily,
getting serious about IPv6 and avoidance of highly centralized
conversion and address translation functions.

    john



<Prev in Thread] Current Thread [Next in Thread>