At 05.44 06/02/02 -0600, you wrote:
>Marco A. Calamari:
> > If normal node doesn't allow key deletion there is one less
> > possible way to exploit a bug.
>
>We don't allow any random host to send FCP commands.
>
>Doesn't that fact negate your (already weak) argument that a
>ClientDelete implementation might not be secure?

No.

Maybe the fact that I'm not (clearly) English mothertongue
  disallow me to clearly explain my point.

IMHO the security model of Freenet is robust *also*
  because negate the existence of a key delete.

I'm not concerned at all about the effect that ClientDelete
  can have when correctly & securely implemented & used.

I fear the possibility that ClientDelete can be exploited
  on a population of well behaved nodes in a well
  designed attempt of general damage.

You cannot exploit only the existing command, not
  those that doesn't exist. One less command, one less
  command exploitable in Fred.

This is not a weak argument about the implementation,
  but an attempt to evaluate a cost/benefit of a new
  command versus the robustness of the Freenet security model.

Please. don't start a thread about the difficulty of find
  an exploit that control a node via FCP interface, this
  is not the point, and we can move on a infosecurity list
  or use private mail if you want discuss existing,
  non-trivial exploit.

I'm not one of developers, in the Project I'm only
  a translator, and maybe an evangelist, nor I want allow me to
  suggest to GJ dropping a part of his excellent job.

I just want to contribute a different point of view
  (I'm a litte more than naive in info security!) to
  a loosely-coupled developer community that now can have more
  focus on work in progress and less on an overall
  aspect of security, *implementation* included.

FWIW.   Marco


* Marco A. Calamari  marco at freenetproject.org *

il  Progetto Freenet - segui il coniglio bianco
the Freenet  Project - follow the  white rabbit


_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to