botnets: web servers, end-systems and Vint Cerf [LONG, sorry]

michael.dillon at bt.com michael.dillon at bt.com
Mon Feb 19 20:36:32 UTC 2007


 
> > Now, even those people have shifted to a hierarchical 
> architecture of
> > instant-messaging servers.
> 
> In what way is IM hierarchial? 
> Jabber/XMPP has a mesh-of-stars topology 

That is hierarchy. One level is a star topology, the next level is a
mesh.

> which is the same as email's
> modulo some simplifications (mainly owing to the lack of forwarding).

In other words, it is not the same as email's. Of course it may end up
that way but we can hope.

> ISTR that you were arguing in favour of a chain-of-trust 
> system for email
> back in November on the IETF list. I pointed out that the 
> architecture you
> are proposing is essentially the same as inter-domain routing 
> (IP & BGP)
> and Usenet, and you failed to explain how your ideas would solve the
> unwanted traffic problem for email given that the same architecture
> doesn't solve the unwanted traffic problem for IP or NNTP.

An abstract simplification of an architecture is not equal to the
architecture itself. The fact that you can simplify different
architectures into a similar abstract model, doesn't mean that they have
the same problems. Problems often arise in the details of
implementation, not in the theoretical models. I never claimed that my
proposed email model would solve the unwanted mail problem. It was
intended to carry authenticated sender info to the receiver, and to
provide an authenticated reverse path for complaints to postmaster. And
since it was based on negotiated bilateral email peering agreements, if
the chain of trust was subverted at some point in the chain, the peer
would have legal recourse to cut service.

--Michael Dillon




More information about the NANOG mailing list