cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2371
Views
0
Helpful
9
Replies

Multicast issues

chrishauser2
Level 1
Level 1

Hi All,

I have an issue regarding Multicast.

We recently upgraded a control system network using 3750G in a "ring" topology, which was running several types of traffic (CCTV:roughly 100 cameras, SOS telephones etc).  Where we have been trying to introduce VLANs, and ran into a few problems due to the CCTV running multicast (was informed it was unicast) and therefore the VLAN introduction has been put on hold until the multicast routing has been resolved.

I have tested it using PIM dense mode but there was a significant delay/latency that arose after a random period of time.  I believe this maybe due to the poor IP address scheme (hardcoded in the control system) which may have cause the problems, such as the following:

  • Multicast packet loops / black holed packets
  • An excess mcast routing table 

Basically I need to know that if I use sparse mode with auto RP will this help!!

Also any advice on this subject would be of great benefit....

If any further details are required dont hesitate to ask

Cheers

Chris

1 Accepted Solution

Accepted Solutions

Chris

What you say actually makes sense. It occured to me after posting that the periodic flooding that occurs with dense mode could be causing your issues so i think it is a good idea to try sparse mode.

Jon

View solution in original post

9 Replies 9

Jon Marshall
Hall of Fame
Hall of Fame

chrishauser2 wrote:

Hi All,

I have an issue regarding Multicast.

We recently upgraded a control system network using 3750G in a "ring" topology, which was running several types of traffic (CCTV:roughly 100 cameras, SOS telephones etc).  Where we have been trying to introduce VLANs, and ran into a few problems due to the CCTV running multicast (was informed it was unicast) and therefore the VLAN introduction has been put on hold until the multicast routing has been resolved.

I have tested it using PIM dense mode but there was a significant delay/latency that arose after a random period of time.  I believe this maybe due to the poor IP address scheme (hardcoded in the control system) which may have cause the problems, such as the following:

  • Multicast packet loops / black holed packets
  • An excess mcast routing table 

Basically I need to know that if I use sparse mode with auto RP will this help!!

Also any advice on this subject would be of great benefit....

If any further details are required dont hesitate to ask

Cheers

Chris

Chris

Were the swtches showing any signs of CPU hit when you get the performance problems ?

Have you enabled IGMP snooping on the switches to stop multicast traffic being flooded to all ports.

Unless your switches are getting overloaded because of the multicast traffic i'm not sure how PIM Sparse mode will help that much. It is not necessarily any quicker to join a group with Sparse mode as it is with Dense mode as far as i am aware and in fact it might actually be a bit slower.

Jon

Hi Jon,

Thanks for the speedy reply!

I didnt notice any High CPU alarms coming in at the time....

I didnt enable IGMP snooping, but I thought it was enable by default?(not sure now tho!)

My reasoning for sparse mode that then all the multicast is send to RP switch, maybe it would help if I give brief overview of the topology (see attached).  All the multicast receivers are located on one switch, the cameras are located thoughout the network.  I think (not confident) the problem was that the routing table was getting overloaded (due to PIM dense mode) but also due to the poor IP addressing.  So my thinking behind PIM sparse mode was to set the 172.17.1.1 switch as the RP then the multicast would be forced there as this is effectively where all the receivers are anyway??

Another prob is that the cameras arent always being viewed and therefore liable to be pruned off, this could also be an issue that I forgot to mention....

I am open to better ideas however so thanks for your help and I will look into it!!

Cheers

Chris

Chris

What you say actually makes sense. It occured to me after posting that the periodic flooding that occurs with dense mode could be causing your issues so i think it is a good idea to try sparse mode.

Jon

Hi,

I set it up in spase-dense mode with a RP, this fixed the issues

Cheers

Giuseppe Larosa
Hall of Fame
Hall of Fame

Hello Chris,

if there are 100 cameras each of them is a multicast traffic source!

is each camera using a different multicast address or the same on all?

Quite a number, you may want to consider bidirectional PIM or source specific multicast.

Sparse mode requires explicit joins, dense mode sends everywhere and requires explicit prune.

Using sparse mode with autoRP may help, but probably bidirectional PIM can be the best choice.

if all cameras send on the same multicast address you should use source specific multicast PIM SSM.

Hope to help

Giuseppe

Hi Giuseppe,

No each camera has its own multicast address that is related to its unicast address e.g 172.16.1.161 would be 239.16.1.161.....

Thanks for the help I will look into Bidirectional PIM

Kind Regards

Hi Giuseppe,

Can you think of any other way to do this other than the above?

Cheers

Chris

Hello Chris,

have you managed to move to sparse mode?

I agree that this is the first step as you and Jon have discussed in previous posts.

See the bidir PIM as a successive step to be used only if multiple cameras are monitored at the same time.

Hope to help

Giuseppe

Hi Giuseppe,

I have not made the change yet as it is part of a live tunnel...

The plan D Day is 15th of January.....

I let you know how I got on!!

Cheers

Chris

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco