Hi Gabriel,
Please see the comments below.

> -----Original Message-----
> From: gabriel montenegro [mailto:[EMAIL PROTECTED]
> Sent: Friday, October 14, 2005 3:31 AM
> To: Ki-Hyung Kim; 'Samita Chakrabarti'; [EMAIL PROTECTED]
> Cc: [email protected]
> Subject: RE: [6lowpan] format document issues
> 
> thanks, comment below...
> 
> --- Ki-Hyung Kim <[EMAIL PROTECTED]> wrote:
> > > Q2: Did you use the handling rules as per the draft, or did you modify
> the
> > > link-layer source address at every hop, as implied by your post a
> while
> > back?
> > I didn't change anything of the format document on the simulation. Also
> the
> > link-layer source address is changed at every hop as everybody expected.
> 
> As I explained before, this departs from the handling rules in the format
> draft.
> According to those rules (I won't send them again), the source address
> does not
> change hop-by-hop. The assumption is that if an intermediate node wants to
> send an
> RERR to its previous hop, it would do a reverse forwarding decision. To be
> clear,
> if multiple paths were kept, presumably an intermediat node would have
> more than one
> reverse path to the originator of a message. In this case, it is not clear
> to which of
> the possible predecessors should the RERR be sent. Perhaps to all of them,
> which costs
> a bit more.
> 
> You say that the "link-layer source address is changed at every hop as
> everybody
> expected".
> 
> One conclusion is that given that this contradicts the current handling
> rules, then it is
> 
> clear that they need to be clarified by explicitly mentioning that no such
> thing is done.
> 
> Another possible conclusion of your "as everybody expected" is that
> between these
> choices:
> 
> a) link-layer source address is changed at every hop
> b) link-layer source address is not changed at every hop
> 
> the draft has obvioiusly made the wrong choice in going with "b". You have
> verified that
> having only one link-layer source address (the one in the 802.15.4 header)
> works under handling rule "a". Have you also evaluated under handling rule
> "b" in order
> to compare
> both? Or is "b" obviously wrong for some reason?

I have looked at your description about the handling rules of the format
right before.
(Sorry, I had no time to look at the mailing list in these days.)
I believe that (a) is right because it is MAC address. I have also double
checked it by the packet sniffer of IEEE802.15.4. That is, (b) is not
correct.



> 
> -gabriel


_______________________________________________
6lowpan mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/6lowpan

Reply via email to