[IP] Re: en of story F.B.I. Gained Unauthorized Access to E-Mail

[David Farber <dave@farber.net>]

________________________________________
From: Steven M. Bellovin [smb@cs.columbia.edu]
Sent: Monday, February 18, 2008 10:05 AM
To: David Farber
Cc: Suresh Ramasubramanian
Subject: Re: [IP] en of story  F.B.I. Gained Unauthorized Access to E-Mail

On Mon, 18 Feb 2008 06:26:50 -0800
David Farber <dave@farber.net> wrote:

>
> There is nothing at all here that suggests that this subpoena was
> unwarranted, a warrantless wiretap etc.

I agree -- that's what I wrote in my blog entry
(http://127.0.0.1/blog/2008-02/2008-02-16.html).
>
> A technical error at an ISP, the FBI notices and asks the ISP to stop.
> Destroys the rest of the data (if not all of it .. there's enough
> rules of evidence to make out a case that whatever was gathered has
> been tainted). End of story.
>
No, I think there's more.  The lesson is that mistakes like this do
happen, and proper *system* design takes that into account.  That might
be mundane -- have a procedure in place for dealing with unauthorized
data, rather than trusting individual agents.  (I have vague
recollections of reading a few years ago about some Carnivore
intercepts being destroyed by an agent who was very upset about
overcollection.)  The more important point, though, is to take this as
a lesson in humility in system design.  This time, a filtering error
caused some emails to be delivered improperly.  Next time, might a
filtering error in a surveillance system lead to penetration of
communications carrier networks?  Before you say "that can't happen",
think of the Greek cellphone scandal.  This is precisely why I and
others have long opposed tight integration of surveillance systems into
communications networks.  It's not that wiretaps are bad; it's that
such mechanisms are all too easily subverted.


                --Steve Bellovin, http://www.cs.columbia.edu/~smb

-------------------------------------------