I have an IF/RF station that is receiving RF signals from satellites, then converted to IP to be encoded, decoded, multiplexed, and modulated again to RF so that it's sent to wireless and wired network to reach the receivers to watch satellite TV.

The station has a network of L3 Cisco switches configured with ip pim dense mode under each vlan that is participating in a multicast stream in order to connect the streams' multicast group addresses to the devices that is re-generating the streams.

The question here is that I believe that we should move from PIM dense mode to sparse mode but I need a good reason to convince the management to do that since they are saying we don't need because all the devices are requesting all the streams always in any case so there's no need to convert it to sparse mode.

Any ideas about if in such cases where the devices requesting the streams are densely populated on all the vlans it's still recommended to switch to sparse mode?

Note that we have only L3 switches and no routers at all. And IP multicast routing is enabled on all the switches.

  • 3
    There is no right answer. Sparse mode is usually preferred over dense, but I'm not sure it's worth converting, unless you are having (bandwidth) problems. Why do you want to change (other than just because)? – Ron Trunk Dec 20 '19 at 13:41
  • 2
    Also, it would be helpful to know how many switches and vlans you have. – Ron Trunk Dec 20 '19 at 13:41
  • 2
    Dense mode usually wastes more bandwidth, so scaling expectations are worse. Switching over would make sense when bandwidth needs to be preserved and you know that all streams and subscribers also work in sparse mode. In some scenarios it wouldn't make much of a difference, so you'll need to add a lot more information to your question to the nature of the network, protocols, streams and subscription. – Zac67 Dec 20 '19 at 15:22
  • 1
    Suffered an outage back in the day (~2002) due to dense mode causing replicated multicast packets (our mode was sparse-dense but the traffic was dense-related). If I was deploying a new multicast network today I would definitely go sparse. But as your manager, given the business critical nature of your multicast, I would be very reluctant to change to sparse unless dense was definitely broken. “If it’s not broken don’t fix it.” – Darrell Root Dec 23 '19 at 5:45

The first point is that sparse mode is inherently more efficient in most networks, as multicast traffic is only sent upon a station actually joining a given group. In contrast, dense mode sends all multicast traffic to all networks until a lack of a querier causes it to time out and start pruning.

In general dense mode has been deprecated in actual use for a while and, indeed, isn't even supported on a pretty broad swath of Cisco switching and routing - specifically anything running NXOS, ACI or IOS-XR. Put another way, you're going to have an increasingly difficult time universally supporting dense mode on a lot of newer hardware. All that said, the Catalyst line (running IOS-XE) still has support for dense and sparse-dense, but in actual practice it's pretty much a legacy compatibility feature.

| improve this answer | |
  • Thanks for all the comments. Afterall, I guess that I shouldn't do this change since there are no problems with bandwidth or outage. We are broadcasting channels using UDP streams and 99% all channels multicast IPs are being used at all the times since the multiplexers are getting these IPs as inputs to output the stream to the QAM modulators, so going to sparse might change nothing although Cisco recommends it in similar cases. – Ali Zaghir Dec 28 '19 at 16:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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