Re: [Autoconf] Multicast confusion between MANET Arch and AUTOCONFProblem Statement

Velt, R. (Ronald) in 't wrote:
> Hi Alex,
> 
>> -----Original Message----- From: Alexandru Petrescu
>> [mailto:alexandru.petrescu@gmail.com] Sent: dinsdag 20 november
>> 2007 16:20 To: autoconf@ietf.org Subject: [Autoconf] Multicast
>> confusion between MANET Arch and AUTOCONFProblem Statement
>> 
>> I think there may be a confusion with respect to the interpretation
>> of multicast in draft-ietf-autoconf-manetarch-07.txt vs 
>> draft-ietf-autoconf-statement-02.txt.
>> 
>> It seems as if the manetarch accepts that link-layers have a well
>> defined multicast behaviour whereas the problem statement doesn't.
>> 
>> autoconf-statement:
>>> Traditional solutions assume that a broadcast directly
>> reaches every
>>> router or host on the subnetwork, whereas this generally is not
>>> the case in MANETs (see [2]).
>> So a broadcasted message (a special case of multicast) will not
>> reach every host on the MANET subnet.
> 
> "link" versus "subnetwork" are distinct notions, aren't they? What is
>  left of a "subnetwork" in an environment where \128 (\32 for IPv4) 
> prefixes are configured on MANET interfaces? If your local interface
> is the only one in the subnet, then all interfaces in the subnet are 
> trivially reachable :-) "Link" is something else...

Right... :-)

But this sounds as special cases.

Maybe simple a 1-1 mapping between link and subnet would be easier.
Just have a link and a subnet be the same, and a simple prefix assigned
to it.  And see these problems first.

The MANET I'd build would surely have a prefix for a subnet on a link 
and /128 addresses for each entity.  This should work.

>> [2] autoconf-manetarch:
>>> Link-local Multicast/Broadcast Scope On a MANET interface, a
>>> packet sent to a link-local multicast or all-ones broadcast
>> address reaches
>>> the MANET interfaces of neighboring MANET routers...
>> So it actually does.
>> 
> 
> No, not all interfaces of all MANET routers. Just those of the 
> *neighbo(u)ring* MANET routers. See MANET architecture for a
> definition of neighbo(u)r.

Ok, manetarch:
> Neighbor In the context of routing, two routers are neighbors if one
> can send/receive routing protocol IP packets to the other without 
> passing through any intermediaries at the same layer.

Ok but in the Internet Area a Neighbor is from rfc4861:
> neighbors   - nodes attached to the same link.

This different 'Neighbor' definition is potentially  another source of
confusion when talking to each other.

I think manetarch may mean actually that a Neighbor is an entity within
a close physical distance, whereas an RFC4861 Neighbor is not
necessarily close by.

BEsides, the definition above is unclear because of that "at the same
layer".  Which layer?  IP?  Link?  Would a link-layer device bridging
messages to a remote-distance entity make this latter a Neighbor?

Wouldn't it sound better like: "CloseEntity: two routers are
closeentities if one can send/receive link-layer multicast messages
to/from the other".

If we don't clarify this it will always pop up as a confusing item when
saying 'Neighbor'.

Thanks,

Alex

______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email 
______________________________________________________________________

_______________________________________________
Autoconf mailing list
Autoconf@ietf.org
https://www1.ietf.org/mailman/listinfo/autoconf