Re: When IGMPv3 router in EXCLUDE(X, Y) state receives TO_IN(A)

Hi,

Thanks for the clarification, bu the following doubt still remains.

In that the group specific query alone should be sufficient to pull all
information related to groups in which hosts are currently interested/not
interested. Why send both the queries? Also that the current exclude list
contains Y-A which is basically those sources which should still be blocked.
Only this set of sources could also be queried.

This statement can also be interpreted as if {X-A} is NULL then the group
specific query has to go out (otherwise the query would have got suppressed
due to statement "If source-list A is null as a result of the action (e.g.,
A*B) then no query is sent as a result of the operation."

thanks and regards,
Rohan Sen

On Wed, Mar 12, 2008 at 11:33 PM, Mark Fine <[email protected]>
wrote:

> The Group-Specific query is to find out if there are any hosts which
> still have EXCLUDE state -- the Group-Specific query will lower the
> router's Group Timer to LMQT, and if no hosts respond with IS_EX
> Current-State records, the Group Record will switch to INCLUDE filter-
> mode if there are still Source Records with timers running.
>
> On Mar 12, 2008, at 2:46 AM, Rohan Sen wrote:
>
> > Hi,
> >
> > RFC 3376 Section 6.4.2 change state record handling router state
> > machine last line mentions about sending a group specific query in
> > addition to the source specific query.
> > I am not very sure if I am able to understand the necessity of this
> > group specific query. Usually a TO_IN(A) message is received from a
> > host that underwent a state change from EXCLUDE(A) to INCLUDE(A)
> > which means the router already had a state of EXCLUDE(X,Y) where A
> > is a subset of Y. The router then tries to find out if there is any
> > host which still has interest in receiving traffic for the sources
> > already having source timer > 0. I do not understand what is the use
> > of sending a group specific query in this case since the group
> > specific query has a wider scoped purpose than the source specific
> > query.
> >
> > Could it be that since for other cases no query is sent if sources
> > to be queried is NULL set, for this case only (since TO_IN(NULL) is
> > the IGMPv3 translation for IGMPv2 Leave when older version host for
> > that group is present) a group specific query has to be sent
> > regardless of whether the to be queried source set is NULL or not.
> >
> > Any help on these lines would be highly appreciated.
> >
> > --
> > thanks,
> > Rohan Sen _______________________________________________
> > magma mailing list
> > [email protected]
> > https://www.ietf.org/mailman/listinfo/magma
>
>


-- 
thanks,
Rohan Sen

_______________________________________________
magma mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/magma