Re: What is an IGMPv3 Router's initial current state foramulticast group?

"Alvaro Fernandez" <[email protected]> Fri, 7 Mar 2008 10:22:06 +0100
Newsgroups gmane.ietf.magma
Message-ID <[email protected]>
  

Hi Mark,

 

I am happy to follow this discussion with you.

 

If there is a reset in a router and after the reset the router receives a TO_IN report, if the router considers the initial state after reset is INCLUDE with an empty list then the router will not send the group specific query of row 12 in the previous table I sent before ( the table is based on RFC 3376.)

 

But I think you can not say that because there is nor register in the router there is no problem with that.

 

If this is a problem or not depends on the reason why the router sends the group specific query in the row 12 of the table. Now I don't remember this reason. I studied it months ago and I must check my notes. I will check.

 

Best regards

Alvaro Fernandez

 

Note: I don't know why but from yesterday I am not receiving the e-mails from magma list.  If you send any response, please cc it to me. Thanks.


________________________________

De: Mark Fine [mailto:[email protected]]
Enviado el: jue 06/03/2008 22:15
Para: Alvaro Fernandez
CC: Suvendu Mozumdar; [email protected]
Asunto: Re: [magma] What is an IGMPv3 Router's initial current state foramulticast group?




> Right, except if there is a reset in the router (as Suvendu wrote 
> "lets say a router comes up afresh ") before receiving the TO_IN 
> message.
>
> This is an example when considering INCLUDE with an empty list  as 
> intial state for the router has some problems.
>
> But I think is the only case and the only difference is the Group 
> specific query action in the router.
> I don't remeber now why this Group specific query is necessary.

This is not a problem for the router: the difference in the New Router 
State and the difference in the Actions "if there is a reset in the 
router" is because of the non-existent state in the reset router. In 
the reset router there is no Old Router State to transition from, 
there is no source records to lower timers and send Group-and-Source-
Specific Queries for, and there is no group record to lower timer and 
send Group-Specific Query for.

Orthogonally, routers use a reduced Query Interval on startup to 
minimize the duration of unlearned Router State when "there is a reset 
in the router", see 8.6 and 8.7.

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