[nylug-talk] 0wned .gov machines (was Re: Russian cyberwar against Estonia?) (fwd)

[Jay Sulzberger]

---------- Forwarded message ----------
  Date: Sun, 20 May 2007 09:03:08 -0600
  From: Anne & Lynn Wheeler <lynn at garlic.com>
  To: "[UTF-8] Ivan KrstiÄ~G" <krstic at solarsail.hcs.harvard.edu>
  Cc: Perry E. Metzger <perry at piermont.com>,
      "Trei, Peter" <ptrei at rsasecurity.com>, cryptography at metzdowd.com
  Subject: Re: 0wned .gov machines (was Re: Russian cyberwar against Estonia?)

  Ivan Krstić wrote:
  > I think it's anything but surprising. There's only so much you can do to
  > significantly improve systems security if you're unwilling to break
  > backwards compatibility -- many of the fundamental premises of desktop
  > security are fatally flawed, chief among them the idea that all programs
  > execute with the full privileges of the executing user.

  part of this is that many of the basic platforms providing internet
  connectivity
  evolved from disconnected/unconnected desk/table top environment ... with
  lots of applications assuming that they had full & free access to all
  resources.

  attempting to leverage the same platforms for connectivity to extremely
  hostility
  and anarchy of the internet creates diametrically opposing requirements.

  one countermeasure from the 60s is to use a dynamically created ("padded cell")
  virtual machine for internet connectivity ... with limited scope and accesses.
  then when the session completes ... the environment is collapsed and everything
  is discarded.
  while the "native" system operation may have little or no defenses against the
  hostile internet ... the "padded cell" virtual machine environment is used to
  bound the scope of any penetration ... somewhat analogous to "air gapping".

  recent post:
  http://www.garlic.com/~lynn/2007k.html#48

  somewhat older reference:
  http://www.nsa.gov/selinux/list-archive/0409/8362.cfm

  ---------------------------------------------------------------------
  The Cryptography Mailing List
  Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com