Re: [Roll] [roll] #5: DODAG
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Roll] [roll] #5: DODAG
Hi Pascal, all,
> -----Original Message-----
> From: Pascal Thubert (pthubert)
> Sent: lundi 9 novembre 2009 12:18
> To: Julien Abeille (jabeille); Tim Winter; ROLL WG
> Subject: RE: [Roll] [roll] #5: DODAG
>
> Hi Julien:
>
> In classical terms, an instance really defines a DAG.
If you have several roots with several DAG IDs, the DAGs are totally
disconnected. This is not one DAG, but several.
The differences we are talking about is "one vs several", not one kind
versus another DAG.
> So if
> you have multiple roots then RPL partitions that DAG into a
> set of Destination-Oriented DAGs, also called DAG Rooted at a
> Destination in the art. All those terms exist with a specific
> load and the wisdom we hear from Phil and Dave is not to
> overload / redefine / restrict those terms.
All DAGs in RPL are destination oriented. I don't see the point of
adding as many adjectives as possible to the DAG term, especially since
we've been talking about DAGs since two years.
>
> The most useful object is the DODAG iteration. A node moves
> within one and jump between 2. This is the object for which
> we should find cool name. DODAGI is certainly pronounceable
> by many of us, but it looks a lot like a WG name in mobility.
> Any idea?
DAG ;-)
Best,
Julien
>
> Cheers,
>
> Pascal
>
> >-----Original Message-----
> >From: roll-bounces at ietf.org [mailto:roll-bounces at ietf.org]
> On Behalf Of
> >Julien Abeille (jabeille)
> >Sent: lundi 9 novembre 2009 11:12
> >To: Tim Winter; ROLL WG
> >Subject: Re: [Roll] [roll] #5: DODAG
> >
> >Hi Tim,
> >
> >My question is: are there DAGs that are not DODAGs in RPL?
> My feeling
> >is no. Hence I would suppress the DODAG instead of turning all "DAG"
> >into "DODAG".
> >Best,
> >Julien
> >
> >> -----Original Message-----
> >> From: roll-bounces at ietf.org [mailto:roll-bounces at ietf.org]
> On Behalf
> >> Of Tim Winter
> >> Sent: samedi 7 novembre 2009 19:55
> >> To: ROLL WG
> >> Subject: [Roll] [roll] #5: DODAG
> >>
> >> Hi Julien, WG,
> >>
> >>
> >> roll issue tracker wrote:
> >> > #5: DODAG
> >> >
> >>
> --------------------------------+------------------------------------
> >> --------------------------------+-
> >> > --------------------------------+------
> >> > Reporter: jpv at ... | Owner: wintert at ...
> >> > Type: defect | Status: new
> >> > Priority: major | Milestone:
> >> > Component: rpl | Version:
> >> > Severity: Active WG Document | Keywords:
> >> >
> >>
> --------------------------------+------------------------------------
> >> --------------------------------+-
> >> > --------------------------------+------
> >> > * This is currently being discussed by the RPL Author team *
> >> >
> >> > Email from Julien
> >> >
> >> > Is there a distinction between DAG and destination
> >> oriented DAG. From
> >> > what I read I feel all DAGs are destination oriented.
> If this is
> >> > correct I propose to remove the destination oriented
> DAG concept.
> >> >
> >>
> >>
> >> The intent for -05 will be to make the unqualifed use of
> `DAG' in the
> >> text consistent with `DODAG Iteration', as outlined below. So in
> >> this use, a `DAG' is a `DODAG'
> >> snapshotted by a single sequence number -- a `DODAG Iteration'.
> >> There may be cases in -04 where the text reads `DAG' and
> is in fact
> >> referring to `DODAG'- we will be sure to clean these up.
> Here below
> >> is some additional explanation prepared by the author team-
> >>
> >> Does this help to clarify?
> >>
> >> Regards,
> >>
> >> -Tim
> >>
> >>
> >>
> >> Instance
> >>
> >>
> >>
> >> +----------------------------------------------------------------+
> >> |
> >> |
> >> | (R1) (R2)
> >> (Rn) |
> >> | / \ /| \ / |
> >> \ |
> >> | / \ / | \ / |
> >> \ |
> >> | (A) (B) (M) | (N) ... (X) (Y)
> >> (Z) |
> >> | /|\ |\ / | |\ | |
> >> | |
> >> | : : : : : : (O) : : : :
> >> : |
> >> | / \
> >> |
> >> | : :
> >> |
> >> |
> >> |
> >> | Instance ID
> >> |
> >> |
> >> |
> >>
> >> +----------------------------------------------------------------+
> >>
> >> Instance
> >>
> >>
> >> An Instance is a routing topology over an LLN, optimized for a
> >> particular objective/application. Discrete Instances
> may also be
> >> set
> >> up to offer optimized connectivity to different
> destinations when
> >> appropriate, for example to differentiate a Home
> Network LLN from a
> >> Utility Network LLN in a smart metering application
> where a meter
> >> may
> >> be configured to join both Instances.
> >>
> >> It consists of one or more Destination Oriented DAGs (DODAGs).
> >>
> >> It is uniquely identified by an InstanceID.
> >>
> >> Each instance is operated independently of other
> instances, and RPL
> >> defines operation over only one instance. Operation
> among multiple
> >> instances is to be expanded upon in a future revision
> or companion
> >> specification.
> >>
> >>
> >>
> >> Destination Oriented DAG (DODAG)
> >>
> >>
> >> +----------------+
> >> | |
> >> | (R1) | (R2) (Rn)
> >> | / \ | /| \ / | \
> >> | / \ | / | \ / | \
> >> | (A) (B) | (M) | (N) ... (X) (Y) (Z)
> >> | /|\ |\ | / | |\ | | |
> >> | : : : : : | : (O) : : : : :
> >> | | / \
> >> | DAGID | : :
> >> | |
> >> +----------------+
> >>
> >> DODAG
> >>
> >>
> >> A Destination Oriented DAG is a DAG rooted at a single
> root node,
> >> which is a node with no outgoing edges.
> >>
> >> In the case where multiple nodes in the LLN coordinate over a
> >> backbone and expose the same DAGID (to support the same
> DODAG), it
> >> is
> >> conceptually as if there is a single virtual root over
> the backbone
> >> (not illustrated). In some
> applications/implementations this may
> >> be
> >> a desired architecture; in other applications each DODAG may
> >> operate
> >> with indpendent uncoordinated roots exposing different DAGIDs.
> >>
> >> A DODAG is uniquely identified over the LLN by the tuple
> >> (InstanceID,
> >> DAGID).
> >>
> >> In RPL a node may belong to only one DODAG per Instance
> at a time.
> >>
> >>
> >>
> >> DODAG Iteration
> >>
> >>
> >> +----------------+ +-----------------+
> >> | | | |
> >> | (R1) | | (R1) |
> >> | / \ | \ | / | \ |
> >> | / \ | ------\ | / | \ |
> >> | (A) (B) | \ | (A) (C) (B) |
> >> | /|\ |\ | / | /|\ |\ |
> >> | : : : : : | ------/ | : : : : : |
> >> | | / | |
> >> | Sequence N | | Sequence N+1 |
> >> | | | |
> >> +----------------+ +-----------------+
> >>
> >> DODAG Iteration
> >>
> >>
> >> A DODAG Iteration is the DODAG that results from the iterative
> >> process that reshapes the DODAG as stimulated by the
> root. It is a
> >> DAG as constrained by operation of RPL over a fixed Sequence
> >> Number.
> >>
> >> As the root node increments the Sequence Number,
> different types of
> >> node movement are allowed (e.g. moving `down') and a new DODAG
> >> Iteration is formed.
> >>
> >> A DODAG Iteration is uniquely identified over the LLN
> by the tuple
> >> (InstanceID, DAGID, SequenceNumber).
> >>
> >>
> >>
> >> Scope
> >>
> >> o The scope of an InstanceID is the whole network and
> it defines
> >> an
> >> instance
> >>
> >> o The scope of a DAGID is an instance and it defines a DODAG
> >> within
> >> that instance
> >>
> >> o The scope of a Sequence Number is a DODAG and it defines an
> >> iteration of that DODAG (DODAG Iteration)
> >>
> >> o The scope of a rank is a DODAG Iteration and it defines a
> >> position
> >> of a node within that iteration.
> >> _______________________________________________
> >> Roll mailing list
> >> Roll at ietf.org
> >> https://www.ietf.org/mailman/listinfo/roll
> >>
> >_______________________________________________
> >Roll mailing list
> >Roll at ietf.org
> >https://www.ietf.org/mailman/listinfo/roll
>
Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.
