Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I'm introducing some PIM routing into my network to make some multicast traffic available across routers.

I have read a bit on PIM, sparse mode versus dense mode. After reading, I am trying to figure out, under what circumstances would I want to use dense mode?

If I have a simple structure:

[multicast subscriber] <-> [DR router] <-> [DR router] <-> [RP router] <-> [multicast sender]

is it true that using sparse mode from this topology would allow me to expand more efficiently?

How is "time to first packet received" affected by sparse versus dense mode, considering a particular large tree of routers? Are RPs always subscribed to receive multicast?

Thanks,

Matt

share|improve this question

1 Answer 1

up vote 3 down vote accepted

The main difference between SM and DM is that DM floods initially until pruned, and SM does not forward until joined. The initial flooding caused by DM is often over stated. The initial flooding will stop within seconds and the optimal MC path used by DM and SM will end up to be the same.

DM has the advantage of being simpler to configure and debug. Also in terms of time to first packet received, DM will also have the advantage because DM floods by default. But the difference between DM and SM in this aspect will be a few seconds at most, unless your network involves hundreds of routers.

RPs will always receive multicast traffic at the start of the flow, but once the optimal path to the destination has been figured out by PIM, RP may no longer receive the multicast traffic if it does not fall within the optimal path.

I hope this helps.

share|improve this answer
    
Thanks for your response. So, the answer is simply: Dense Mode is just a little worse in efficiency, better with "time-to-first-packet", and much easier to configure. So in a cost-benefit, Dense always wins. Then why use Sparse? :) Only in VERY large networks? –  mbrownnyc Feb 11 '13 at 15:09
1  
I've only come across one situation where I would have configured SM over DM. In this network there were so many multicast streams that if any client joined all streams at once then network congestion would occur. This would be one good reason to us PIM-SM, because if all the streams started at the same time (after a server reboot or something), under PIM-DM, the initial flooding would cause network congestion that could lead to further problems. –  wookie919 Feb 11 '13 at 23:09
1  
To clarify the above example, regardless of whether DM or SM is used, congestion would have occurred if a client started listening on all streams. It is only the initial flooding of all multicast streams that can be prevented With PIM-SM (and this is a good enough reason to use SM over DM). –  wookie919 Feb 11 '13 at 23:15
    
Thanks for the additional replies! –  mbrownnyc Feb 14 '13 at 16:03

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.