Re: [manet] Encoding an IPv6 Mobile Network Prefix with PacketBB

[Thomas Clausen <T.Clausen at computer.org>]

On Jun 19, 2007, at 6:16 PM, Alexandru Petrescu wrote:

> Thomas Clausen wrote:
> > On Jun 19, 2007, at 5:49 PM, Alexandru Petrescu wrote:
> > > Dearlove, Christopher (UK) wrote:
> > > > > I came up with the datagram I present first.
> > > > I'm afraid I don't have time to comment on this.
> > > Thanks, I request MANET community to comment on that PAcketBB
> > > encoding of the IPv6 Mobile Network PRefix - thank you.
> > I still do not think that it is clear what it is you are trying to  
> > encode, and your figure doesn't help. What is it that you are trying
> > to do? In your figure, it seems that you are trying to have part of
> > an address block header and then a TLV block followed by an IPv6  
> > address.....then another TLV block and some padding.
> > Doesn't look to me as this encoding would be compatible with
> > packetbb, but if I understood what it was you were trying to do, then
> > I am sure that it could be encoded.
> > Have you looked at draft-ietf-manet-packetbb (here's an URL:  
> > http://www.ietf.org/internet-drafts/draft-ietf-manet-packetbb-05.txt)
> >  section 5 for the description of the format, and  appendix A for  
> > illustrations? It should provide you with inspiration as to how you
> > can encode things, once they're specified, and examples of how the
> > actual layout would be in the familiar IETF-pictures.
>
> YEs I've looked at these sections.  I have plenty of remarks on  
> them.  I
> refrain from writing remarks about the way that section is written,  
> for
> fearing of risking be not well understood.
>
> All I need is a consensus-based PacketBB message that contains an IPv6
> address with an IPv6 prefix length on it.  Just one, not two.  No
> head-mid-tail and no index-start nor index-stop.

Let me play devils advocate: you want a packetbb message that's not a  
packetbb message?

> I've tried to propose one as I understand it.  Would you please remark
> whether you agree with that message or not and why - thank you.

Ok, I do not agree with that message you proposed. Why? Because it is  
not a validly formed packetbb message.

> > If all you want is a single IPv6 prefix, then you'd benefit little
> > from any aggregation, and a message TLV with explicit content might
> > simply do the trick. But, I am not sure I understand if that is what
> > you are looking for.
>
> YEs, that is what I am looking for: a PacketBB message TLV with  
> explicit
> content, containing an IPv6 Mobile Network Prefix (an IPv6 address and
> the prefix length).

Very well. Seeing that you want that and only that one single prefix  
ever, then I'd go with a message TLV and an explicit representation  
of the prefix. See the examples in appendix A for how such a message  
TLV would look.

However if you might have more prefixes to advertise, or more  
addresses, then this would not be the efficient solution. But, I feel  
that you're either aiming for a very specific situation -- or for  
demonstrating something specific -- so I am not sure to which extend  
it is helpful that I comment on that without knowing the whole design- 
space you are considering.

Please do not take this as not-wanting-to-help -- but as the prudence  
that good engineers have to exhibit in not proposing a solution  
without understanding the problem and the design-space.

Thomas





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