Re: [manet] multi-gateway support draft

[Francisco Ros <fjrm4 at yahoo.es>]

El Viernes, 10 de Septiembre de 2004 18:28, Shubhranshu escribiÃ:
> Francisco,
>
> Its a typo. It should be [src=MN, dst=AR] [dst1=H]
>
Sorry for disturbing again :-), but if C.GW sends such a packet then H will 
not be able to answer with [src=H, dst=AR] [dst1=C.GW, dst2= MN] (because it 
won't know about existence of C.GW).

By the way, is there any advantage by using proxy routing header? If you don't 
use it there exists a simpler solution:
MN -> [src=MN, dst= C.GW] [dst1=H]
H -> [src=H, dst=C.GW] [dst1= MN]

What do you think about this? Thanks in advance.

Regards, Francisco Ros

>
> ----- Original Message -----
> From: "Francisco Ros" <fjrm4 at yahoo.es>
> To: "Shubhranshu" <shubhranshu at samsung.com>
> Cc: <manet at ietf.org>
> Sent: Thursday, September 09, 2004 7:50 AM
> Subject: Re: [manet] multi-gateway support draft
>
>
> Thank you very much for your answers, now I can understand it better :-).
> But I am still having a doubt. Please see below.
>
> El Viernes, 10 de Septiembre de 2004 07:43, Shubhranshu escribiÃ:
> > ----- Original Message -----
> > From: "Francisco Ros" <fjrm4 at yahoo.es>
> > To: <manet at ietf.org>
> > Sent: Wednesday, September 08, 2004 7:33 AM
> > Subject: [manet] multi-gateway support draft
> >
> > > Hi everybody!
> > >
> > > I am studying several proposals for providing Internet connectivity to
> > > manets. Now I am reading "Mobile multi-gateway support for IPv6 mobile
> > > ad hoc networks" (draft-singh-manet-mmg-00) and have some doubts
> > > regarding it.
> > >
> > > I cannot understand properly section 4.2 (Routing policy). Please look
> > > at this two paragraphs:
> > >
> > > <<
> > >    Default gateway can route traffic originating from manet node and
> > >    whose destination is internet node using default route while the
> > >    candidate gateways should route such traffic either using proxy
> > >    routing header or tunneling mechanism.  In proxy routing header
> > >    destination address of IPv6 header is set to the Internet access
> > >    router address while final destination address set to internet node
> > >    address and original source address set to routing header.  Proxy
> > >    routing header indicates to internet node that the packet travel via
> > >    candidate internet gateway and access router.  In response to this
> > >    received packet, internet node must use candidate internet gateway
> > >    and original source address in the routing header.
> > >
> > >
> > > As far as I understand, ongoing packets look like [src=MN, dst1=AR,
> > > dst2=H] and incoming ones like [src=H, dst1=CG, dst2=MN]. This has no
> > > sense for me so I suppose I am misunderstanding something, aren't I?
> > > (MN = manet node, AR = access router, H = internet host, CG = candidate
> > > gateway)
> > >
> > > Besides specification does not mention how packets sent by MN are. I
> > > think they should by like [src=MN, dst1=CG, dst2=H].
> > >
> > > <<
> > >    If manet node uses tunneling mechanism, candidate internet gateways
> > >    encapsulate packets to the internet node.  Tunneling header has
> > >    source address as candidate internet gateway and destination address
> > >    as internet node.  The internet node decapsulates the packet and
> > >    replies with tunneling header with source address as internet node's
> > >    address and destination address as candidate internet gateway's
> > >    address.
> > >
> > >
> > > Could someone explain how are packets sent by MN, CG and H please?
> > > I suppose that they could be (alhough is not clear for me):
> > > MN -> [src=MN, dst1=CG, dst2=H]
> > > CG -> [src=CG, dst=H][src=MN, dst=H]
> > > H -> [src=H, dst=CG][src=H, dst=MN]
> > >
> > > I hope that someone could clarify all this. Thanks in advance!
> >
> > Two possible cases:
> >
> >  First, mobile node uses the routing header (considering it knows AR
> > address). In this case it should be [src=MN, dst=C.GW] [ dst1=AR, dst2=H]
> >
> >  Second, candiadate gateway uses the proxy routing header. In this case
> > it receives packet from MN with [src=MN, dst= C.GW] [dst1=H]  then the
> > candidate gateway changes it as  [src=MN, dst=AR] [dst1=H, dst= C.GW]
>
>                                                                            
>        ^^^^^^^^^^^^^ Is this a typo or did you actually want to write that?
> As far as I know this would cause the remote host to forward the _same_
> packet to C.GW, which is not the desired behavior (at least if we are using
> routing headers of type 0 as defined by RFC 2460). Are you thinking about
> some special type of routing headers to support correct behavior? Or what
> am I missing?
>
> Thanks again ;)
>
> Regards, Francisco Ros
>
> >  After receiving packet with routing header, internet node replies to MN
> > with header:  [src=H, dst=AR] [dst1=C.GW, dst2= MN]
> >
> >  ( C.GW=candidate gateway address )
> >
> >  New draft version with better/detail explanation will be published soon.
> >
> > Thanks
> > Shubhranshu
>
> _______________________________________________
> manet mailing list
> manet at ietf.org
> https://www1.ietf.org/mailman/listinfo/manet
>
>
>
> _______________________________________________
> manet mailing list
> manet at ietf.org
> https://www1.ietf.org/mailman/listinfo/manet

_______________________________________________
manet mailing list
manet at ietf.org
https://www1.ietf.org/mailman/listinfo/manet