RE: Simple dual homing experiment

[Coene Lode <Lode.Coene@siemens.com>]

[ post by non-subscriber.  with the massive amount of spam, it is easy to miss
  and therefore delete posts by non-subscribers.  if you wish to regularly
  post from an address that is not subscribed to this mailing list, send a
  message to <listname>-owner@ops.ietf.org and ask to have the alternate
  address added to the list of addresses from which submissions are
  automatically accepted. ]

> -----Original Message-----
> From: Christian Huitema [mailto:huitema@windows.microsoft.com]
> Sent: Thursday, March 20, 2003 3:03 AM
> To: kurtis@kurtis.pp.se
> Cc: multi6@ops.ietf.org
> Subject: FW: Simple dual homing experiment
> 
> 
> Kurtis,
> 
> David and I wrote up a "simple dual homing" proposal, based 
> on previous drafts and private conversations. This may be one 
> of several ways of going forward in multi6.
> 
> -- Christian Huitema
> 
> 
>
issue 3.1 : egress filtering:

I would think that selecting the destination address  based on the source
address is a bit strange...
Transport protocols(SCTP in particular) selects a destination address, the
IP layer selects an outgoing interface, thus in order to have a msg with the
correct source address to get through egress filtering, the msg should
contain the IP address of the outgoing interface as this is the address the
interface got assigned by the provider upstream(and the provider would let
its own sourceaddresses/prefixes through, wouldn't it..??..)

issue 3.2:
some form of heatbeating per destination(as in SCTP) should do the trick...
no problem there

issue 3.3:
looks the same from the perspective of the transport layer.. use same
solutions as in 3.2
solve one, get a extra free with...

issue 3.4:
changeover to a active connection/path.. some transport protocols can do
that also.. no problem..

issue 3.5:
actually the only guy that could do optimisations is the endpoint itself: it
has the congestion information, it could know the load distribution between
the different paths/connections if she is smart enough.. 

basically the smart solution should not pose any problems.. atn least it
hasn't give us(=SCTP implementations) any problems up till now. And multiple
TCP connections(or something in that direction) would deliver the same
stuff...

yours sincerly,
Lode Coene

Siemens atea
atealaan 34          2200 Herentals, Belgium
E-mail: lode.coene@siemens.com
Tel: +32-14-252081
Fax: +32-14-253212