Orhan Ergun 2 Comments

One of my students asked me this question a month ago. “What is the difference between PIM Sparse Mode and PIM SSM (Source Specific Multicast)?”

But, since I had two CCDE bootcamps in one month, I didn’t have time to answer the question on this platform.

By the way I have seen this mis understanding in the recent popular design books as well, so writing about it was necessary anymore.

Okay, let me give you the answer – an answer specially dedicated to those who started to say, what are you really comparing!

They are the same thing.

In order to understand why, let’s define PIM Sparse Mode and PIM Dense Mode.

In multicast network, there is a multicast sender and receiver. If the sender sends the multicast stream everywhere (which mean each multicast-enabled interface is on the network without waiting for other interface), it is a PIM Dense Mode. In fact, that’s why it is called push based multicast mechanism.

When multicast stream is sent to each link and if there is no intended receiver from that sender to the multicast group, the routers will send prune message to stop the stream for a while.

On the other hand, if the receivers signal their intent to receive multicast stream to the centralized node, it is called PIM Sparse Mode. So, multicast stream is not sent everywhere in PIM Sparse Mode. That’s why it is known as pull based multicast mechanism.

There are 3 different PIM Sparse Mode deployment models: PIM ASM (Any Source Multicast), PIM SSM (Source Specific Multicast), and PIM Bidir (Bidirectional PIM).

The centralized node in PIM ASM and PIM Bidir, both of which are used by the receiver to send their intent for multicast traffic, is called PIM Rendezvous Point (PIM RP)

In PIM SSM, centralized node – which is used by multicast receiver to signal their intent for multicast traffic – is called multicast sender or multicast source. So, there is no rendezvous point in PIM SSM (Source Specific Multicast)

That’s why the difference between PIM Dense and PIM Sparse is this: In PIM Dense Mode, the receiver doesn’t signal its intention for multicast traffic, but in PIM Sparse mode the receiver initiates the operation. Put simply, the receiver signals whether or not it is interested in multicast traffic.

Since the multicast receivers start the whole operation by sending their intent to either the RP or the multicast source, PIM ASM, PIM SSM, PIM Bidir are called PIM Sparse Mode as well.

I compared PIM SSM, PIM ASM and PIM Bidir in a comprehensive comparison table which helps network engineers to decide which solution is best for their network.

 
0.00 avg. rating (0% score) - 0 votes
  • Grzegorz Wypych

    Orhan I will also mention the choice between Optimal routing vs Optimal resources in the mcast network:

    PIM-SSM – because routers holds (S,G) entry for each active multicast source it best fit to network with low number of sources and high number of receivers. Each additional multicast source requires new (S,G) entry to be created. In the other hand because Root of the tree is always source traffic is optimal (base on IGP metrics for example)

    PIM-SM – needs 2 trees to function, both rooted on RP (Source and shared tree). RP hold (*,G) entires which star represents any multicast source in the network. That means if we will have Many sources for the same multicast group only single (*,G) entry is required on RP per mcast group which provides more optimal resource allocation. The cost is optimal traffic flow because by default all traffic is received on RP and forwarded by RP (possible to join SPT tree for optimal traffic from receiver to source, but even after that for each new mcast group RP is forwarding point of the information about active source).

    PIM-Bidir – Bidirectional trees offer are more optimal than unidirectional PIM-SM and they are able forward data in both directions. There is no source tree which also provides optimal resource allocation. (good for apps when source could also be receiver)

    • Orhan Ergun

      Hi Greg, Exactly. Aggregated state might cause sub-optimal routing but in case of failure, faster convergence.

      Aggregation will provide better resource usage thus scalability, but stretch of the multicast flow might be an issue.

      I covered almost all of the other aspects in the PIM Sparse Mode article which you may wanna take a look at too