RE: Teredo vs Silkroad

["liumin" <liumin@ict.ac.cn>]

> -----Original Message-----
> From: owner-v6ops@ops.ietf.org [mailto:owner-v6ops@ops.ietf.org] On
Behalf
> Of JF Tremblay
> Sent: Saturday, May 22, 2004 12:04 AM
> To: rengrong wang; v6ops@ops.ietf.org
> Subject: Re: Teredo vs Silkroad
> 
> --On May 21, 2004 2:31 PM +0800 rengrong wang <rengronw@usc.edu>
wrote:
> 
> > Hi,
> >
> > In draft-liumin-v6ops-silkroad-01.txt it is said that Silkroad wants
to
> > enable nodes located behind one or several IPv4 NATs to obtain IPv6
> > connectivity and it seems like a tunnel-broker solution.  It is
known
> > that Teredo is a automatic tunnel mechanism that figures out the
same
> > problem.
> >
> > What's the difference between Silkroad and Teredo?
> 
> IMHO, Silkroad is simply another protocol based on tunnel
broker/server
> model (RFC3053) with an optimization to make two hosts on the same
link
> realize they can talk directly. Teredo on the other hand allows hosts
> implementing a Teredo client to exchange directly through NATs, except
> symmetric ones. However traffic to hosts not implementing a Teredo
client
> must go through a relay.

I don't think so. TSP is a tunnel broker solution, which is not
presented to solve NAT transversal but to provide one easy way to
configure and maintain many different tunnels. From this point, TSP has
relation to all existing tunnel technologies. But it is a "pure
procedure" management technology for tunnels. It is totally different
from tunnel technologies which focus on methodology for tunnels.  

Silkroad is not an implementation of tunnel broker like TSP.

 

> 
> At first sight, Silkroad doesn't seem to offer any significant
improvement
> over TSP, except the nodes-on-same-link optimization. Both support NAT
> traversal with any type of NAT, including nested ones. However TSP
doesn't
> require the use of a web page, offers prefix delegation,
authentication and
> has a solid experimental background since it's been deployed for over
5
> years to more than 100000 users.

Web pages is just one way for Silkroad client to learn its SAR, there
are many other ways to reach this point, including XML presentation.


> 
> The question sparking in my mind from this is whether of not the
detection
> of two tunneled hosts on the same link would be a desirable feature to
> include in the tunnel broker model. It involves the broker telling a
host
> what is the public IPv4 address and port of another host, which may be
a
> security issue. However with Teredo this information is already
included in
> the IPv6 address, so I guess it would provide the same level of
security.
> 
> > Best regards,
> >
> > Crisy(Rengrong) Wang
> > =======================================
> > USC,EE-Systems
> >
> >
> 
> Jean-Francois Tremblay
> Hexago
> -------------------------------------------------
> http://www.freenet6.net : Free IPv6 Connectivity
> -------------------------------------------------
> "Computer Science is no more about computers than
> astronomy is about telescopes" - E. W. Dijkstra
> -------------------------------------------------
>