Skip to Content.
Sympa Menu

bluesky - Re: The Grapevine Project

bluesky AT lists.ibiblio.org

Subject: Global-Scale Distributed Storage Systems

List archive

Chronological Thread  
  • From: Stephen Blackheath <stephen AT blacksapphire.com>
  • To: "Global-Scale Distributed Storage Systems" <bluesky AT franklin.oit.unc.edu>
  • Subject: Re: The Grapevine Project
  • Date: Thu, 6 Dec 2001 16:14:26 +1300


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'll answer several things in one post

On Thu, 06 Dec 2001 19:19, Tony Kimball wrote:
> Another commodity which is typically used is human cognitive capacity.
> A network of many participants has a lot of attached wetware while
> even the most resourceful attacker can typically muster only a
> relatively small troop of human agents.

That's a good thought.

On Thu, 06 Dec 2001 14:43, Steve Jenson wrote:
> Stephen Blackheath wrote:
> > On Thu, 06 Dec 2001 13:00, you wrote:
> >>Stephen Blackheath wrote:
> >>>We use a 'hash cash' technique
> >>>(which we call "Solve A Hard Problem") for defence against
> >>>denial-of-service attacks - the assumption being that legitimate nodes
> >>>have more CPU power than attackers.
> >>
> >>that seems like a silly assumption. What about DDoS?
> >
> > Firstly, understand that when I say "legitimate nodes have more CPU
> > power", I am referring to the sum of all legitimate nodes vs. the sum of
> > all attackers.
>
> So it's not so much an assumption as a constraint? I'm not trying to be
> nitpicky but to clear up ambiguity. Speaking of constraints, I'm unable
> to find a clear, concise Threat Model established in your literature.

It's more of a construmption, really.

That's because there isn't a clear concise Threat Model in my literature.
However - there is a really shitty security analysis near the end of
'grapevine.pdf', which is a little out of date now.

> > Ultimately, if an attacker has more of something than you do, then the
> > attacker will win. The only thing you have control over as designer is
> > what that something is. We chose CPU time rather than network bandwidth
> > or access to IP addresses, because none of the strategies relating to the
> > latter things seemed to work. I am certainly open to better ideas.
>
> This isn't necessarily true, have you read Ross Anderson's excellent
> work 'Security Engineering' (ISBN 0471389226)? He outlays possible
> strategies to protect agains *DoS, one of which is late binding of
> names.

I'm not familiar with that - thanks. I'll take a look.

> > DDoS - if you're talking about it as an attack from nodes outside the
> > network - - is difficult against the Grapevine due to the high cost of
> > 'harvesting' IP addresses.
>
> No, I'm talking about malicious nodes from within your network working
> in tandem.

OK. I won't go into detail here, but the SAHP strategy is the answer to the
bulk of the various classes of denial-of-service attacks.

On Thu, 06 Dec 2001 15:14, Wes Felter wrote:
> on 12/5/01 6:20 PM, Anthony Jones at ajones AT clear.net.nz wrote:
> > Just like any application, there will be disk space requirements. They
> > will not necessarily be large but they will not necessarily be
> > adjustable.
>
> To put it a different way, what will a Grapevine node do when it gets a
> "disk full" error from the OS?

That's a good question. We haven't really resolved that yet.


Steve
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8DuKT7I0ehz47OHERApcOAJ43O+UbrTKxtR4Q1I9jTX+NY2npKgCg15qp
SH4kpV89c2P8+IaXisb3Zb4=
=DCsh
-----END PGP SIGNATURE-----




Archive powered by MHonArc 2.6.24.

Top of Page