Re: [Roll] [roll] #145: The same value for PROACTIVE_FORWARDING should be not recommended

["roll issue tracker" <[e-mail]>]

#145: The same value for PROACTIVE_FORWARDING should be not recommended


Comment (by [e-mail]):

 Thread: http://www.ietf.org/mail-archive/web/roll/current/msg08506.html

 From: Robert Cragie <robert.cragie at gridmerge.com>
 Date: Wed, 12 Feb 2014 13:31:50 +0000

 Isn't PROACTIVE_FORWARDING now a parameter per MPL Interface (based on the
 contested statement mentioned below, i.e. "It is RECOMMENDED that all MPL
 Interfaces ... use the same value for PROACTIVE)? Therefore section 5.4
 should just be SEED_SET_ENTRY_LIFETIME and there should be a new section
 "MPL Interface Parameters"

 Other comments inline.

 Robert

 On 10/02/2014 2:03 PM, Jonathan Hui (johui) wrote:
 Hi Axel,

 >Thanks for your review and comments.  See my comments below:

 >>On Feb 8, 2014, at 1:26 AM, Axel Colin de Verdière <axel-
 [e-mail]> wrote:

 >>I've reviewed the draft, and I have a couple comments/questions if
 that's not too late:

 >>- I don't see why MPL forwarders on the same link should have the same
 value for >>PROACTIVE_FORWARDING. It is now RECOMMENDED to make it so, but
 I don't see the rationale for this >>recommendation: on the opposite, I
 think it would allow some forwarders with less power >>restrictions to be
 more proactive than others.
 >Good point.  If there are no objections, I can remove this RECOMMENDED
 statement.
 <RCC>
 I agree it is not strictly necessary but there are certain combinations of
 parameters which won't work. For example, if proactive forwarding is used
 and the CONTROL_MESSAGE_TIMER_EXPIRATIONS is 0 then a MPL Forwarder's MPL
 Interface which doesn't use proactive forwarding may never forward a MPL
 Data Message on that interface as none of its neighboring MPL Forwarders
 will be sending MPL Control Messages.

 I know Kerry had some concerns earlier regarding mutual exclusivity of
 proactive and reactive forwarding strategies. Introduce the wider scope
 through multiple interfaces and thus per-MPL Interface parameter sets
 makes sense but it brings in a new concept of grouping a strategy across a
 common set of MPL Interfaces to ensure forwarding applies correctly on
 those MPL Interfaces.
 </RCC>

-- 
-------------------------------------+-------------------------------------
 Reporter:                           |       Owner:  draft-ietf-roll-
  [e-mail]          |  [e-mail]
     Type:  defect                   |      Status:  new
 Priority:  major                    |   Milestone:
Component:  trickle-mcast            |     Version:
 Severity:  In WG Last Call          |  Resolution:
 Keywords:                           |
-------------------------------------+-------------------------------------

Ticket URL: <http://trac.tools.ietf.org/wg/roll/trac/ticket/145#comment:2>
roll <http://tools.ietf.org/wg/roll/>