Re: [6lowpan] [Roll] Fwd: New Version Notification for draft-goyal-6lowpan-rpl-compression-00.txt

[Mukul Goyal <[e-mail]>]

>As a distant observer of the ROLL WG, one thing I don't understand is why this is necessary in the first place.
>As far as I understand, the domain of RPL is low-power (i.e., constrained) networks.
>6LoWPAN is just one of the network types RPL will be used on, and redoing this work for every other type (that benefits from >compactness) sounds wasteful to me.
>Why doesn't RPL itself define a reasonably compact representation then?

I think that is a good idea. The reasons I targetted 6lowpan:

1) In my view, RPL over 6lowpan is a common case.
2) 6lowpan already has a mechanism for next header compression, that we could use.

Thanks
Mukul 

----- Original Message -----
From: "Carsten Bormann" <[e-mail]>
To: "Mukul Goyal" <[e-mail]>
Cc: "6lowpan" <[e-mail]>, "roll" <[e-mail]>
Sent: Thursday, May 19, 2011 11:26:54 AM
Subject: Re: [Roll] Fwd: New Version Notification for draft-goyal-6lowpan-rpl-compression-00.txt

Mukul,

I just looked at your draft in a little more detail.
Certainly, a purpose-built spec to squeeze out redundancy from a specific packet format will generally be more efficient than the generic compression I have written up.

However, the arguments in section 1.1 of draft-bormann-6lowpan-ghc-02.txt apply to the fullest here.
Since the HC spec needs to be understood by all nodes, it creates a powerful obstacle in evolving the subject protocol.  The strong coupling between the subject protocol and the compression spec also creates opportunity for interesting interoperability problems.

As a distant observer of the ROLL WG, one thing I don't understand is why this is necessary in the first place.
As far as I understand, the domain of RPL is low-power (i.e., constrained) networks.
6LoWPAN is just one of the network types RPL will be used on, and redoing this work for every other type (that benefits from compactness) sounds wasteful to me.
Why doesn't RPL itself define a reasonably compact representation then?

Gruesse, Carsten