Re: [Roll] draft-ietf-roll-trickle-mcast-03: Varying Trickle parameters per MPL Domain

[Abdussalam Baryun <[e-mail]>]

Hi Esko,

comments in line,

On 1/30/13, Dijk, Esko <[e-mail]> wrote:
> In Section 5.4, Trickle parameters are (implicitly?) forced to be equal for
> all MPL Domains that an MPL Forwarder participates in, because only 2 sets
> are supported per MPL Forwarder. Isn't this overly restrictive? Given that
> the MPL Domains don't interact with each other, it would be fine if a set of
> parameters is equal only within one and the same MPL Domain (so, that means
> an MPL Forwarder could have different Trickle parameters for each MPL Domain
> it is in.)

I think having more than one domain per forwarder makes success of
multicast complicated. If different trickel parameters for domains,
the drafts should explain more about this issue,
>
> The final sentence of section 5.4 confirms this view, that only
> within-Domain parameter equality is important:
> "It is RECOMMENDED that all MPL Forwarder within an MPL Domain use the same
> values for the Trickle Parameters above"
>

So it is recommended, so it can be different in one domain, which
gives flexibility to the protocol in one domain,

> Since a configuration method for Trickle parameters is out of scope of the
> I-D, it doesn't add any extra complexity here to allow different Trickle
> parameters for different MPL Domains. Also, allowing this does not seem to
> break the protocol in any way (unless I overlook something.)

You cannot just make every thing out of scope to avoid complexity. If
the protocol allows same forwarder with different domains, I think it
is important at least to identify a parameter for different domains or
explain how forwards, then its config can be out of scope, but just
mentioning that the protocol can allow multi-domains per forwarder, I
think there should be an answer of how forwarder does such activity,

My thoughts. I may be missing something because not completed my review,

AB