Re: [pim] RFC 4541

Hi Bill,

Trying to understand passive listeners from a protocol interaction point of view. Is there an example of (2) passive listener case from IANA assigned multicast groups in this range ?

PS: Thanks for correcting. Sorry, I actually meant "assigned" by "reserved" in my email. 

regs
Ganesh

-----Original Message-----
From: William Atwood <william.atwood@concordia.ca> 
Sent: Friday, November 9, 2018 2:03 AM
To: ChennimalaiSankaran, Ganesh
Cc: pim@ietf.org
Subject: Re: [pim] RFC 4541 - 224.0.0.*

[EXTERNAL EMAIL] 
Please report any suspicious attachments, links, or requests for sensitive information.

Hi,

Three things:

1) The "reserved" numbers in the range 224.0.0.x are as yet unassigned.
No protocol should be using them.  In the future, IANA may be requested
to make a new assignment in this range; this will normally happen as the
result of the issuing of an RFC (describing, for example, a new
protocol).  So one has to assume the same properties in this range as
for all presently know (i.e., assigned) protocols.

2) NO protocol operating in the range 224.0.0.x need ever issue an
explicit join.  As you quote in your first message, other than IGMP, all
 switches MUST forward every packet received to every port on the
switch.  This is _precisely_ because there may be silent listeners on
that port.

3) Concerning IGMP and MLD, I suspect that if you look elsewhere, you
will find that IGMP v2 packets and MLD v1 packets must be forwarded in
the same way, but IGMP v3 and MLD v2 packets need not be.  (This is
because the latest versions of IGMP and MLD have different semantics
from the previous versions.)

  Bill

On 08/11/2018 11:51 AM, Ganesh.ChennimalaiSa@Dell.com wrote:
> Hi Stig,
> 
> Going through the IANA registry, I see the below class of protocols 
> 1. bidirectional protocols such as OSPF, RIPv2 etc
> 2. protocols used to locate a service such as mDNS etc
> 3. reserved but usage not known
> 
> First two are chatty but as you said the response can be a unicast/multicast. Unfortunately, protocols under the third class does not have public documentation.
> 
> regs
> Ganesh
> 
> 
> -----Original Message-----
> From: Stig Venaas <stig@venaas.com> 
> Sent: Thursday, November 8, 2018 2:40 PM
> To: ChennimalaiSankaran, Ganesh
> Cc: pim@ietf.org
> Subject: Re: [pim] RFC 4541 - 224.0.0.*
> 
> 
> [EXTERNAL EMAIL] 
> Please report any suspicious attachments, links, or requests for sensitive information.
> 
> 
> Hi
> 
> OK, I think I understand. Isn't it a problem if a protocol expects to
> be able to receive before sending though? If you look at
> https://www.iana.org/assignments/multicast-addresses/multicast-addresses.xhtml#multicast-addresses-1
> you will see that there are quite a few different protocols, it is
> difficult to say how all of them behave. Also, we don't know how
> future protocols might behave.
> 
> Also, I hope you aren't assuming that a router/host that wants to
> receive for a specific group, would be sending messages to the same
> group address. There could be client-server link-local protocols that
> don't work that way.
> 
> Stig
> 
> On Thu, Nov 8, 2018 at 12:40 AM <Ganesh.ChennimalaiSa@dell.com> wrote:
>>
>> Hi Stig,
>>
>> I understand the point. My question is: when all protocols are chatty we theoretically can use an alternate mechanism to learn their existence and forward only to group members. For instance OSPF is chatty and OSPF PDUs can be forwarded only within OSPF routers by leveraging its chatty nature. But when we have passive listeners in a link-local group, there is no opportunity for the network to know about the existence of such receivers.
>>
>> Do we have passive receivers (or protocols) in this 224.0.0.* range ?
>>
>> regs
>> Ganesh
>>
>> -----Original Message-----
>> From: pim <pim-bounces@ietf.org> On Behalf Of Stig Venaas
>> Sent: Thursday, November 8, 2018 1:58 PM
>> To: ChennimalaiSankaran, Ganesh
>> Cc: pim@ietf.org
>> Subject: Re: [pim] RFC 4541 - 224.0.0.*
>>
>>
>> [EXTERNAL EMAIL]
>> Please report any suspicious attachments, links, or requests for sensitive information.
>>
>>
>> Hi
>>
>> The problem is that many stacks don't send IGMP reports for link-local
>> groups, making the assumption that only routers need to know about
>> membership, and link-local groups are not routed. Hence, you cannot
>> rely on IGMP to determine which ports have interested receivers.
>>
>> Stig
>>
>> On Sun, Nov 4, 2018 at 4:55 PM <Ganesh.ChennimalaiSa@dell.com> wrote:
>>>
>>> Hi,
>>>
>>>
>>>
>>> Looking at RFC4541 and specifically the discussion 2..1.2 (2) on data forwarding rules. It says
>>>
>>>
>>>
>>> “Packets with a destination IP (DIP) address in the 224.0.0.X range
>>>
>>>       which are not IGMP must be forwarded on all ports.”
>>>
>>>
>>>
>>> As I see, there may not be explicit joins in this range and looking at IANA registry most protocols look chatty.
>>>
>>>
>>>
>>> Are there any groups with silent listeners that use this range ?
>>>
>>>
>>>
>>> regs
>>>
>>> Ganesh
>>>
>>>
>>>
>>> _______________________________________________
>>> pim mailing list
>>> pim@ietf.org
>>> https://www.ietf.org/mailman/listinfo/pim
>>
>> _______________________________________________
>> pim mailing list
>> pim@ietf.org
>> https://www.ietf.org/mailman/listinfo/pim
> _______________________________________________
> pim mailing list
> pim@ietf.org
> https://www.ietf.org/mailman/listinfo/pim
> 

-- 

Dr. J.W. Atwood, Eng.             tel:   +1 (514) 848-2424 x3046
Distinguished Professor Emeritus  fax:   +1 (514) 848-2830
Department of Computer Science
   and Software Engineering
Concordia University EV 3.185     email:william.atwood@concordia.ca
1455 de Maisonneuve Blvd. West    http://users.encs.concordia.ca/~bill
Montreal, Quebec Canada H3G 1M8