Re: netscape's response

[Frank A Stevenson <frank@funcom.no>]

On Wed, 20 Sep 1995, Jeff Weinstein wrote:

> NOTE: my first attempt to send this bounced at toad.com
> 
> On Sep 20,  5:16pm, David_A Wagner wrote:
> > Subject: Re: netscape's response
> > In article <9509200139.ZM206@tofuhut> you write:
> > > On Sep 20, 12:29am, Christian Wettergren wrote:
> > > > One wild idea that I just got was to have servers and clients exchange
> > > > random numbers (not seeds of course), in a kind of chaining way. Since
> > > > most viewers connect to a number of servers, and all servers are
> > > > connected to by many clients, they would mix "randomness sources" with
> > > > each other, making it impossible to observe the local environment
> > > > only. And the random values would of course be encrypted under the
> > > > session key, making it impossible to "watch the wire".
> > > 
> > >   Wow, this is a great idea!!
> > 
> > Are you quite sure this is a good idea?
> > 
> > I'd be very scared of it.  In particular, it opens up the chance for
> > adversaries to feed you specially chosen numbers to pollute your seeds.

Suppose you divide your random material into several parts:

A: Userinput (updated from Keystroke timing etc.)
B: 'Random' numbers from remote server
C: Time, pid, ppid, etc..
D: other...

Whenever you want to incorporate new data into B you could do something like:
B = B xor Hash (A,B,C,D, fresh 'random')

This would be very hard to pollute with well chosen input.

> 
>   What I should have said is that its a very interesting idea.  Given
> current perceptions of netscape, I should have made clear that I
> wouldn't do something like this without getting a lot more discussion
> and review of possible dangers and how to avoid them.  I certainly
> can't fault anyone for wondering if we would just implement this
> without thinking it through, given recent events.
> 

  Frank