- Information
- AI Chat
Was this document helpful?
Protocol Independent Multicast (PIM)
Course: Systems Programming (01:198:214)
23 Documents
Students shared 23 documents in this course
University: Rutgers University
Was this document helpful?
Protocol Independent Multicast (PIM)
PIM, or Protocol Independent Multicast, was created in response to older multicast routing
technologies' scaling issues. It was discovered, in particular, that the previous protocols did not
scale effectively in situations where just a small percentage of routers desire to receive data for a
specific group. For instance, if the majority of routers don't wish to receive the traffic in the first
place, broadcasting traffic to all routers until they expressly request to be removed from the
distribution is not a wise design decision. Because of how frequent this occurrence is, PIM splits
the problem space into sparse mode and dense mode, where sparse and dense denote the
percentage of routers that will require the multicast. Similar to DVMRP, PIM dense mode (PIM-
DM) uses a flood-and-prune method and has scaling issues.
We will concentrate on PIM sparse mode (PIM-SM), which has supplanted other multicast
routing protocols. By the way, the "protocol independent" component of PIM refers to the fact
that, unlike earlier protocols like DVMRP, it can be used with any unicast routing protocol, as we
will see later. PIM does not depend on any specific kind of unicast routing. Using PIM protocol
messages referred to as Join messages, routers actively join the multicast distribution tree in
PIM-SM. Take note of how this differs from DVMRP's strategy, which entails first building a
broadcast tree before trimming the routers who aren't interested in it. It then becomes a matter of
where to send those Join messages as any host (and any number of hosts) are capable of sending
to the multicast group.
To fix this, PIM-SM gives each group access to a unique router called the rendezvous point (RP).
PIM-SM offers a series of methods through which all the routers in a domain can agree on which
router to use as the RP for a specific group. Typically, a number of routers in a domain are
configured to be candidate RPs. These procedures must handle a wide range of situations,
including the failure of a candidate RP and the division of a domain into two independent
networks as a result of several link or node failures, making them highly complex. For the
remainder of this explanation, we'll assume that every router in a domain is aware of the RP's
unicast IP address for each group. As a result of routers sending Join messages to the RP, a
multicast forwarding tree is created. PIM-SM enables the creation of two different types of trees:
a shared tree that may be used by all senders and a source-specific tree that can only be used by a
single transmitting host. If there is sufficient traffic, one or more source-specific trees may then
be created after the shared tree in the regular mode of operation.
It is crucial that there is just one tree by default for a group, not one for each sender to a group, as
creating trees installs state in the routers along the tree. A router uses standard IP unicast
transmission while sending a Join message to the RP for a group G. Figure 4.14(a) serves as an
illustration of this, showing router R4 sending a Join to a group's rendezvous point. The first Join
message is "wildcarded," which means that it is applicable to every sender. Before reaching the
RP (for example, R2), a Join message must undoubtedly pass through a number of routers. Each
router along the path examines the Join and creates a (*, G) entry for the shared tree (where *
stands for "all senders").
