All Cisco-Network Study Notes: 01/02/09

Security

Security 689
Over the years, the CCIE program has expanded to add several CCIE certifications besides the
Routing and Switching track. As a result, some topics previously covered in the Routing and
Switching exam have been removed, or shortened, because they are more appropriate for
another CCIE track. For example, the CCIE Routing and Switching track formerly covered
voice to some degree, but the CCIE Voice track now covers voice to a much deeper level.
The topics in this chapter are certainly covered in more detail in the CCIE Security written and
lab exams. However, because security has such an important role in networks, and because
many security features relate specifically to router and switch operations, some security details
remain within the CCIE Routing and Switching track. This chapter covers many of the core
security features related to routers and switches.

RFC Reference-mroute Flags

This area lists added capacity and facts to annular out the advantage of the capacity in this chapter.

Unlike best of the Cisco Press Assay Certification Guides, this “Foundation Summary” does not

repeat advice presented in the “Foundation Topics” area of the chapter. Please booty the

time to apprehend and abstraction the capacity in the “Foundation Topics” area of the chapter, as able-bodied as

review items acclaimed with a Key Topic icon.

Table 17-5 lists the agreement standards referenced in this chapter.

Table 17-6 lists some of the best accustomed Cisco IOS commands accompanying to the capacity in this

chapter and Affiliate 16.

Table 17-5 RFC Reference for Affiliate 20

RFC What It Defines

3973 PIM-DM

3618 MSDP

3446 Anycast RP

2362 PIM-SM

1584 Multicast Extensions to OSPF

Table 17-6 Command Reference for Chapters 16 and 17

Command Command Approach and Description

ip multicast-routing Global mode; appropriate aboriginal command on Cisco routers to use

multicasting.

ip pim dense-mode1 Interface config mode; configures the interface to use PIM-DM

routing protocol.

ip pim sparse-mode1 Interface config mode; configures the interface to use PIM-SM

routing protocol.

ip pim sparse-dense-mode Interface config mode; configures the interface to use PIM-SM

routing agreement for a accumulation if the RP abode is known;

otherwise, uses PIM-DM acquisition protocol.

Foundation Arbitrary 635

Command Command Approach and Description

ip igmp adaptation {1 | 2} Interface config mode; sets the IGMP adaptation on an interface. The

default is 2.

ip igmp query-interval abnormal Interface config mode; changes the breach for IGMP queries

sent by the router from the absence 60 seconds.

ip igmp query-max-responsetime

seconds

Interface config mode; changes the Max Response Time

advertised in IGMP Queries from the absence of 10 abnormal for

IGMPv2 and IGMPv3.

ip igmp join-group group-address Interface config mode; configures a router to accompany a multicast

group. The group-address is a multicast IP abode in four-part

dotted-decimal notation.

ip multicast abuttals access-list

[filter-autorp]

Interface config mode; configures an interface as a multicast

boundary for authoritative scoping. A numbered or named

access account controls the ambit of accumulation addresses afflicted by the

boundary. (Optional) filter-autorp filters Auto-RP messages

denied by the abuttals ACL.

ip multicast ttl-threshold ttl-value Interface config mode; configures an interface as a multicast

boundary for TTL scoping. Time-to-Live amount represents

number of hops, alignment from 0 to 255. The absence amount is 0,

which agency that all multicast packets are forwarded out the

interface.

ip cgmp Interface config mode; enables abutment for CGMP on an

interface.

ip pim adaptation {1 | 2} Interface config mode; sets the PIM adaptation on an interface. The

default is 2.

ip pim query-interval abnormal Interface config mode; changes the breach for PIMv2 Hello or

PIMv1 Router Query letters from the absence 60 seconds.

ip pim message-interval abnormal Interface config mode; changes the breach for sparse-mode Join/

Prune letters from the absence 60 seconds.

ip pim spt-threshold {kbps |

infinity} [group-list access-listnumber]

Global mode; specifies the admission amount for the multicast traffic

for a PIM-SM router to about-face from RPT to SPT. The absence is to

switch afterwards the aboriginal multicast packet is received. If the group-list

option is used, the command ambit are activated alone to the

groups acceptable by the admission list; otherwise, they are activated to

all groups.

Table 17-6 Command Reference for Chapters 16 and 17 (Continued)

continues

636 Affiliate 17: IP Multicast Routing

Command Command Approach and Description

ip pim rp-address rp-address

[access-list] [override]

Global mode; statically configures the IP abode of an

RP area rp-address is a unicast IP abode in four-part,

dotted notation. (Optional) access-list represents a cardinal or

name of an admission account that defines for which multicast groups

the RP should be used. (Optional) override indicates that if

there is a conflict, the RP configured with this command

prevails over the RP abstruse dynamically by Auto-RP or any

other method.

ip pim send-rp-announce

interface-type interface-number

scope ttl-value [group-list accesslist]

[interval seconds]

Global mode; configures the router to be an RP, and the router

sends RP-Announce letters application the Auto-RP adjustment for the

interface abode selected. Ambit represents the TTL. (Optional)

group-list defines the multicast groups for which this router is

RP. (Optional) breach changes the advertisement frequency

from the absence 60 seconds.

ip pim send-rp-discovery

[interface-type interface-number]

scope ttl-value

Global mode; configures the router to be a mapping agent, and

the router sends RP-Discovery letters application the Auto-RP

method. ambit represents the TTL. (Optional) The IP abode of

the interface defined is acclimated as the antecedent abode for the

messages. The absence is to use the IP abode of the interface on

which the bulletin is beatific as the antecedent address.

ip pim rp-announce-filter rp-list

access-list group-list access-list

Global mode; configures a mapping abettor to clarify RPAnnounce

messages advancing from specific RPs. rp-list accesslist

specifies a cardinal or name of a accustomed admission account that

specifies that this clarify is alone for the RP addresses acceptable in

this ACL. group-list access-list specifies a cardinal or name of a

standard admission account that describes acceptable accumulation addresses.

The clarify defines that alone the accumulation ambit acceptable in the

group-list access-list should be accustomed from the RPAnnouncements

received from the RP addresses acceptable by

the rp-list access-list.

show ip igmp groups [group-name

| group-address | interface-type

interface-number] [detail]

User mode; displays the account of multicast groups for which the

router has anon affiliated accumulation members, abstruse via IGMP.

show ip mroute [group-address |

group-name] [source-address |

source-name] [interface-type

interface-number] [summary]

[count] [active kbps]

User mode; displays the capacity of the IP multicast acquisition table.

Table 17-6 Command Reference for Chapters 16 and 17 (Continued)

Foundation Arbitrary 637

Table 17-7 summarizes important flags displayed in an mroute access aback you use the command

show ip mroute.

Command Command Approach and Description

show ip pim acquaintance [interfacetype

interface-number]

User mode; displays the account of neighbors apparent by PIM.

show ip pim rp [mapping [elected

| in-use] | metric] [rp-address]

User mode; displays the alive RPs associated with multicast

groups.

show ip rpf {source-address |

source-name} [metric]

User mode; displays the advice IP multicasting acquisition uses

to accomplish the RPF check.

clear ip cgmp [interface-type

interface-number]

Enable mode; the router sends a CGMP Leave bulletin and

instructs the switches to bright all accumulation entries they have

cached.

debug ip igmp Enable mode; displays IGMP letters accustomed and sent, and

IGMP-host-related events.

debug ip pim Enable mode; displays PIM letters accustomed and sent, and

PIM-related events.

1When you configure any one of these commands on a LAN interface, IGMPv2 is automatically enabled on the

interface.

Table 17-7 mroute Flags

Flag Description

D (dense) Access is operating in close mode.

S (sparse) Access is operating in dispersed mode.

C (connected) A affiliate of the multicast accumulation is present on the anon connected

interface.

L (local) The router itself is a affiliate of the multicast group.

P (pruned) Route has been pruned.

R (RP-bit set) Indicates that the (S,G) access is pointing against the RP. The RP is about in

a pruned accompaniment forth the aggregate timberline afterwards a after router has switched to

SPT for a accurate source.

F (register flag) Indicates that the software is registering for a multicast source.

Table 17-6 Command Reference for Chapters 16 and 17 (Continued)

continues

638 Affiliate 17: IP Multicast Routing

Memory Builders

The CCIE Acquisition and Switching accounting exam, like all Cisco CCIE accounting exams, covers a fairly

broad set of topics. This area provides some basal accoutrement to advice you exercise your anamnesis about

some of the broader capacity covered in this chapter.

Fill in Key Tables from Memory

Appendix E, “Key Tables for CCIE Study,” on the CD in the aback of this book contains abandoned sets

of some of the key arbitrary tables in anniversary chapter. Print Appendix E, accredit to this chapter’s tables

in it, and ample in the tables from memory. Accredit to Appendix F, “Solutions for Key Tables for CCIE

Study,” on the CD to analysis your answers.

Flag Description

T (SPT-bit set) Indicates that packets accept been accustomed on the shortest-path source

tree.

J (join SPT) This banderole has acceptation alone for sparse-mode groups. For (*,G) entries, the

J banderole indicates that the amount of cartage abounding bottomward the aggregate timberline has

exceeded the SPT-Threshold set for the group. This adding is done already a

second. On Cisco routers, the absence SPT-Threshold amount is 0 kbps. When

the J banderole is set on the (*,G) access and the router has a anon connected

group affiliate denoted by the C flag, the abutting (S,G) packet accustomed bottomward the

shared timberline will activate a about-face over from RPT to SPT for antecedent S and

group G.

For (S,G) entries, the J banderole indicates that the access was created because the

router has switched over from RPT to SPT for the group. Aback the J banderole is

set for the (S,G) entries, the router monitors the cartage amount on SPT and

switches aback to RPT for this antecedent if the cartage amount on the antecedent timberline falls

below the group’s SPT-Threshold for added than 1 minute.

Table 17-7 mroute Flags

Memory Builders 639

Definitions

Next, booty a few moments to address bottomward the definitions for the afterward terms:

dense-mode protocol, RPF check, sparse-mode protocol, RP, multicast scoping, TTL scoping,

administrative scoping, PIM-DM, PIM Hello message, appointed router, source-based

distribution tree, multicast accompaniment information, Join/Prune message, upstream router, downstream

router, Graft message, Graft Ack message, Prune Override, Assert message, DVMRP, MOSPF,

PIM-SM, antecedent DR, antecedent registration, aggregate administration tree, shortest-path timberline switchover,

PIM-SM (S, G) RP-bit Prune, Auto-RP

Refer to the comment to analysis your answers.

Comparison of PIM-DM and PIM-SM

Comparison of PIM-DM and PIM-SM
One of the most confusing parts of the PIM-DM and PIM-SM designs is that it appears that if
sources keep sending, and receivers keep listening, there is no difference between the end results
of the end-user multicast packet flow using these two options. Once PIM-SM completes its more
complicated processes, the routers near the receivers have all joined the SPT to the source, and the
most efficient forwarding paths are used for each (S,G) tree.
Although its underlying operation is a bit more complicated, PIM-SM tends to be the more
popular option today. PIM-SM’s inherent strategy of not forwarding multicasts until hosts request
them makes it more efficient during times of low usage. When the numbers of senders and
receivers increases, PIM-SM quickly moves to use the SPT—the same SPT that would have been
derived using PIM-DM. As such, PIM-SM has become a more popular option for most enterprise
implementations today. It has also become a popular option for interdomain multicast as well.
Table 17-4 summarizes the important features of PIM-DM and PIM-SM.
Table 17-4 Comparison of PIM-DM and PIM-SM
Feature PIM-DM PIM-SM
Destination address for
Version 1 Query messages,
and IP protocol number
224.0.0.2 and 2 224.0.0.2 and 2
Destination address for
Version 2 Hello messages,
and IP protocol number
224.0.0.13 and 103 224.0.0.13 and 103
Default interval for Query
and Hello messages
30 seconds 30 seconds
Default Holdtime for
Versions 1 and 2
90 seconds 90 seconds
Rule for electing a
designated router on a
multiaccess network
Router with the highest IP
address on the subnet
Router with the highest IP address
on the subnet
Main design principle A router automatically receives
the traffic. If it does not want the
traffic, it has to say no (send a
Prune message) to its sender.
Unless a router specifically makes a
request to an RP, it does not receive
multicast traffic.
Feature PIM-DM PIM-SM
SPT or RPT? Uses only SPT First uses RPT and then switches to
SPT
Uses Join/Prune messages? Yes Yes
Uses Graft and Graft-Ack
messages?
Yes No
Uses Prune Override
mechanism?
Yes Yes
Uses Assert message? Yes Yes
Uses RP? No Yes
Uses source registration
process?
No Yes

Bidirectional PIM

PIM-SM works calmly with a almost baby cardinal of multicast senders. However, in cases

with a ample cardinal of senders and receivers, PIM-SM becomes beneath efficient. Bidirectional PIM

addresses this about disability by hardly alteration the rules acclimated by PIM-SM.

To acknowledge bidirectional PIM, a abrupt analysis of PIM-SM’s accustomed operations is useful. While

many variations can occur, the afterward accepted accomplish can be acclimated by PIM-SM:

1. The RP builds a aggregate tree, with itself as the root, for forwarding multicast packets.

2. When a antecedent aboriginal sends multicasts, the router abutting the antecedent assiduously the multicasts to

the RP, encapsulated central a PIM Register message.

3. The RP joins the source-specific timberline for that antecedent by sending a PIM Accompany against that source.

4. Later, the routers absorbed to the aforementioned LANs as the receivers can accelerate a PIM Accompany against the

source to accompany the SPT for that source.

With bidirectional PIM, the aftermost three accomplish in this account are not performed. Bidirectional PIM instead

follows these steps:

1. As with accustomed PIM-SM, the RP builds a aggregate tree, with itself as the root, for forwarding

multicast packets.

2. When a antecedent sends multicasts, the router accepting those multicasts does not use a PIM

back up the timberline against the RP. This action continues for all multicast packets from the source.

3. The RP assiduously the multicasts via the aggregate tree.

4. All packets are forwarded per Accomplish 2 and 3. The RP does not accompany the antecedent timberline for the

source, and the blade routers do not accompany the SPT, either.

The name “bidirectional” comes from Step 2, in which the router abreast the antecedent assiduously packets

back up the timberline against the RP. The added administration in the timberline is acclimated at Step 3, with the RP

forwarding multicasts application the aggregate tree.

Anycast RP with MSDP

IP Multicast Routing
Anycast RP with MSDP
The final tool covered here for finding a router’s RP is called Anycast RP with Multicast Source
Discovery Protocol (MSDP). Anycast RP is actually an implementation feature more than a new
feature with new configuration commands. As will be explained in the upcoming pages, Anycast
RP can actually use static RP configuration, Auto-RP, and BSR.
The key differences between using Anycast RP and using either Auto-RP or BSR relate to how the
redundant RPs are used. The differences are as follows:
■ Without Anycast RP—RP redundancy allows only one router to be the active RP for each
multicast group. Load sharing of the collective work of the RPs is accomplished by using one
RP for some groups and another RP for other groups.
■ With Anycast RP—RP redundancy and load sharing can be achieved with multiple RPs
concurrently acting as the RP for the same group
The way Anycast RP works is to have each RP use the same IP address. The RPs must advertise
this address, typically as a /32 prefix, with its IGP. Then, the other methods of learning an RP—static
configuration, Auto-RP, and BSR—all view the multiple RPs as a single RP. At the end of the
process, any packets sent to “the” RP are routed per IGP routes to the closest RP. Figure 17-22
shows an example of the process.
Figure 17-22 Learning the RP Address with Anycast RP
Figure 17-22 shows a design using two RPs (RP-East and RP-West) along with Auto-RP. The steps
shown in the figure are as follows:
1. Both RPs are configured with 172.16.1.1/32, and configured to use that IP address for RP
functions. In this case, both are configured to be the RP for all multicast groups.
R1
Mapping
Agent
RP-Announce
172.16.1.1,
224.0.0.0/4
2
4
RP-Announce
172.16.1.1,
224.0.0.0/4
1 RP so far. Mapping:
All groups, 172.16.1.1
RP-Discovery
Messages
2
3
RP-West RP-East
R-W1 R-E1
Configuration:
RP Address: 172.16.1.1
Groups: 224.0.0.0/4
Use Auto RP
Loop1: 172.16.1.1/32
1
RP Learned via Auto RP:
172.16.1.1
5 RP Learned via Auto RP:
172.16.1.1
5
Configuration:
RP Address: 172.16.1.1
Groups: 224.0.0.0/4
Use Auto RP
Loop1: 172.16.1.1/32
1
Sparse-Mode Routing Protocols 629
2. Both RPs act as normal for Auto-RP by sending RP-Announce messages to 224.0.1.39.
3. The Auto-RP mapping agent builds its mapping table with a single entry, because it cannot
tell the difference between the two RPs, because both use IP address 172.16.1.1.
4. The Auto-RP mapping agent acts as normal, sending an RP-Discovery message to
224.0.1.40. It includes (in this case) a single mapping entry: all groups map to 172.16.1.1.
5. All the routers, including routers R-W1 and R-E1, learn via Auto-RP that the single RP for
all groups is 172.16.1.1.
The last step described in the list brings the discussion to the main benefit of Anycast RP. At this
point, the core Auto-RP function of advertising the IP address of the RP is complete. Of course,
the IP address exists on two routers in Figure 17-22, but it could be more than that in other designs.
Because of the IGP routes, when routers in the western part of the network (like R-W1) send
packets to the RP at 172.16.1.1, they are actually sending the packets to RP-West. Likewise, when
routers in the eastern part of the network (like R-E1) send packets to the RP (172.16.1.1), they are
actually sending the packets to RP-East. This behavior is only achieved by using the Anycast RP
implementation option beyond simply using Auto-RP.
The two biggest benefits of this design with Anycast RP are as follows:
■ Multiple RPs share the load for a single multicast group.
■ Recovery after a failed RP happens quickly. If an RP fails, multicast traffic is only interrupted
for the amount of time it takes the IGP to converge to point to the other RP sharing the same
IP address.
The design of Anycast RP creates a problem that must be overcome using MSDP. The problem
relates to the fact that each individual RP builds its own shared tree, but any multicast source sends
packets to one of the RPs. For example, Figure 17-23 shows the same network as Figure 17-22,
but now with a multicast source in the western part of the network. The routers in the west side of
the figure receive the packets as distributed by RP-West via its shared tree. However, the routers
in RP-East’s shared tree do not get the packets because RP-East never gets the packet sent by the
server in the west side.
The solution to this problem is for the RPs to tell each other about all known sources by using
MSDP. MSDP allows RPs to send messages to each other, revealing the IP addresses of each
source for each multicast group. In Figure 17-23, RP-West could tell RP-East about the multicast
source for 226.1.1.1 at unicast IP address 172.16.5.5. Then, RP-East can join the SPT of source
172.16.5.5, group 226.1.1.1, just as it would have done if it had received the multicast traffic
directly from 172.16.5.5.
630 Chapter 17: IP Multicast Routing
Figure 17-23 The Anycast RP Problem (Later Solved with MSDP)
Summary: Finding the RP
This section covers the concepts behind four separate methods for finding the RP. Three are
specific configuration features, namely static configuration, Auto-RP, and BSR. The fourth,
Anycast RP, actually uses any of the first three methods, but with the design that includes having
the RPs use the same unicast IP address to achieve better redundancy features. Table 17-3
summarizes the methods of finding the RP with PIM-SM.
Table 17-3 Comparison of Methods of Finding the RP
Method RP Details Mapping Info
Redundant
RP
Support?
Load
Sharing
of One
Group?
Static Simple reference to
unicast IP address.
— No No
Auto-RP Sends RP-Announce to
224.0.1.39; relies on
sparse-dense mode.
Mapping agent sends via
RP-Discovery to
224.0.1.40
Yes No
R1
1 4
3
3
RP-West RP-East
R-W1 R-E1
H2
2
Forward on my Shared
Tree – that doesn’t
include the East!
I’m part of RP-East’s
Shared Tree – I don’t
get the packets!
Scope of RP-West’s
Shared Tree
Scope of RP-East’s
Shared Tree
Source for 226.1.1.1
IP Address 172.16.5.5
Method RP Details Mapping Info
Redundant
RP
Support?
Load
Sharing
of One
Group?
BSR Sends c-RP
advertisements as
unicasts to BSR IP
address; does not need
sparse-dense mode.
Sends bootstrap messages
flooded over non-RPF path
Yes No
Anycast RP Each RP uses identical
IP addresses.
Can use Auto-RP or BSR
normal processes
Yes Yes

Dynamically Finding the RP Application BSR

Cisco provided the proprietary Auto-RP affection to break a brace of specific problems. PIM

Version 2, which came later, provided a altered band-aid to the aforementioned problem, namely the

BootStrap Router (BSR) feature. From a actual accepted perspective, BSR works analogously to

Auto-RP. Anniversary RP sends a bulletin to addition router, which collects the group-to-RP mapping

information. That router again distributes the mapping advice to the PIM routers. However,

any assay of BSR above that akin of detail shows that these two accoutrement do alter in

many ways.

It is accessible to aboriginal accept the abstraction of the bootstrap router, or BSR router, afore thinking

about the RPs. One router acts as BSR, which is agnate to the mapping abettor in Auto-RP. The

BSR receives mapping advice from the RPs, and again it advertises the advice to other

routers. However, there are some specific differences amid the accomplishments of the BSR, and their

implications, and the accomplishments of the Auto-RP mapping agent:

■ The BSR router does not aces the best RP for anniversary multicast group; instead, the BSR router

sends all group-to-RP mapping advice to the added PIM routers central bootstrap messages.

■ PIM routers anniversary apart aces the currently best RP for anniversary multicast accumulation by

running the aforementioned assortment algorithm on the advice in the bootstrap message.

■ The BSR floods the mapping advice in a bootstrap bulletin beatific to the all-PIM-routers

multicast abode (224.0.0.13).

■ The calamity of the bootstrap bulletin does not crave the routers to accept a accepted RP or to

support close mode. (This will be declared in added detail in the abutting few pages.)

Figure 17-21 shows an example, declared next, of how the BSR floods the bootstrap message.

PIMv2 creates specific rules for BSR bootstrap messages, advertence that PIM routers should flood

these messages. PIM-SM routers flood bootstrap letters out all non-RPF interfaces, which in

effect guarantees that at atomic one archetype of the bulletin makes it to every router. Note that this logic

is not abased on a alive dense- or spare-mode implementation. As a result, BSR overcomes

the chicken-and-egg botheration of Auto-RP.

For example, in Figure 17-21, brainstorm that R4’s s1 interface is its RPF interface to ability R2, and

R5’s RPF interface to ability R2 is its s0 interface. So, they anniversary advanced the bootstrap letters at

Step 3 of Figure 17-21. However, because R4 receives the bootstrap bulletin from R5 on one of

R4’s non-RPF interfaces, R4 discards the packet, thereby preventing loops. R5 additionally does not

forward the bootstrap bulletin any added for the aforementioned basal reasons.

Sparse-Mode Routing Protocols 627

Figure 17-21 BSR Calamity Bootstrap Messages

The added important allotment of BSR operation is for anniversary applicant RP (c-RP) to acquaint the BSR

router that it is an RP and to analyze the multicast groups it supports. This allotment of the action with

BSR is simple if you accumulate in apperception the afterward point:

All PIM routers already apperceive the unicast IP abode of the BSR based on the earlier

receipt of bootstrap messages.

So, the c-RPs artlessly accelerate unicast messages, alleged c-RP Advertisements, to the BSR. These c-RP

advertisements accommodate the IP abode acclimated by the c-RP, and the groups it supports.

The BSR affection supports bombastic RPs and bombastic BSRs. As mentioned earlier, the bootstrap

message beatific by the BSR router includes all applicant RPs, with anniversary router application the aforementioned hash

algorithm to aces the currently best RP for anniversary multicast group. The mapping advice can list

multiple RPs that abutment the aforementioned accumulation addresses.

Additionally, assorted BSR routers can be configured. In that case, anniversary applicant BSR (c-BSR)

router sends bootstrap letters that accommodate the antecedence of the BSR router and its IP address. The

highest-priority BSR wins, or if a tie occurs, the accomplished BSR IP abode wins. Then, the winning

BSR, alleged the adopted BSR, continues to accelerate bootstrap messages, while the added BSRs

monitor those messages. If the adopted BSR’s bootstrap letters cease, the bombastic BSRs can

attempt to booty over.

R4 R5

BSR

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

E0 .4

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

10.1.8.0/24

E0 .5

1 Bootstrap 1 Bootstrap

2 Bootstrap

3 Bootstrap

2 Bootstrap

Joining the Aggregate Tree

So far, this area on PIM-SM has explained the ancestry of the allotment process, by which

a router abreast the antecedent of multicast packets registers with the RP. Afore commutual that

discussion, however, the abstraction of the aggregate timberline for a multicast group, additionally alleged the root-path

tree (RPT), charge be explained. As mentioned earlier, PIM-SM initially causes multicasts to be

delivered in a two-step process: first, packets are beatific from the antecedent to the RP, and afresh the RP

forwards the packets to the subnets that accept hosts that charge a archetype of those multicasts. PIM-SM

uses this aggregate timberline in the additional allotment of the process.

The RPT is a tree, with the RP at the root, that defines over which links multicasts should be

forwarded to ability all appropriate routers. One such timberline exists for anniversary multicast accumulation that is

currently alive in the internetwork. So, already the multicast packets beatific by anniversary antecedent are

forwarded to the RP, the RP uses the RPT for that multicast accumulation to actuate area to forward

these packets.

PIM-SM routers collectively actualize the RPT by sending PIM Accompany letters against the RP. In

PIM-SM, multicast cartage is beatific alone to routers that accurately appeal it. PIM-SM routers

request the cartage by abutting the RPT by sending a Accompany against the RP.

PIM-SM routers accept to accelerate a Accompany beneath two conditions:

■ Aback a PIM-SM router receives a PIM Accompany bulletin on any interface added than the interface

used to avenue packets against the RP

■ Aback a PIM-SM router receives an IGMP Membership Report bulletin from a host on a

directly affiliated subnet

Figure 17-14 shows an archetype of the PIM-SM accompany process, application the aforementioned network

as Figure 17-12 but with H1 abutting accumulation 228.8.8.8. The routers acknowledge to the IGMP Accompany by

sending a Accompany against the RP, to become allotment of the aggregate SPT (*,228.8.8.8).

Sparse-Mode Acquisition Protocols 613

Figure 17-14 Creating a Aggregate Timberline for (*,228.8.8.8)

Figure 17-14 shows how H1 causes a aggregate timberline (*,228.8.8.8) to be created, as declared in the

following steps:

1. H1 sends an IGMP Accompany bulletin for the accumulation 228.8.8.8.

2. R4 realizes it now needs to ask the RP to accelerate it packets beatific to 228.8.8.8, so R4 sends a PIM

Join for the aggregate timberline for accumulation 228.8.8.8 against the RP. R4 additionally puts its e0 interface into a

forwarding accompaniment for the RPT for accumulation 228.8.8.8.

3. R4 sends the Accompany to the RP.

4. R5 receives the Accompany on its s1 interface, so R5 puts its s1 interface in a forwarding accompaniment for the

shared timberline (represented by (*,228.8.8.8)). R5 additionally knows it needs to advanced the Accompany toward

the RP.

5. R5 sends the Accompany against the RP.

6. R3, the RP, puts its s0 interface in a forwarding accompaniment for the (*,288.8.8.8) aggregate tree.

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

IGMP Join

228.8.8.8

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

PIM-SM Join

(*, 228.8.8.8)

PIM-SM Join

(*, 228.8.8.8)

• Accomplish e0 forward

for (*, 228.8.8.8)

• Accelerate Accompany out RPF

int. to ability RP

4 • Accomplish s1 forward

for (*, 228.8.8.8)

• Accelerate Accompany out RPF

int. to ability RP

6 • Accomplish s0 forward

for (*, 228.8.8.8)

614 Affiliate 17: IP Multicast Routing

By the end of this process, the RP knows that at atomic one host wants packets beatific to 228.8.8.8. The

RPT for accumulation 228.8.8.8 is formed with R3’s s0 interface, R5’s s1 interface, and R4’s e0 interface.

Completion of the Antecedent Allotment Process

So far in this description of PIM-SM, a antecedent (10.1.1.10) beatific packets to 228.8.8.8, as apparent in

Figure 17-13—but no one cared at the time, so the RP did not advanced the packets. Next, you

learned what happens aback a host does appetite to accept packets, with the routers reacting to create

the RPT for that group. This area completes the adventure by assuming how an RP reacts to a PIM

When the RP receives a Register bulletin for an alive multicast group—in added words, the RP

believes that it should advanced packets beatific to the group—the RP does not accelerate a Register-Stop

message, as was apparent aback in Figure 17-13. Instead, it reacts to the Register bulletin by deencapsulating

the multicast packet, and forwarding it.

The behavior of the RP in acknowledgment to the Register bulletin credibility out the additional above function

of the Register message. Its capital two functions are as follows:

■ To acquiesce a router to acquaint the RP that it has a bounded antecedent for a accurate multicast group

■ To acquiesce a router to advanced multicasts to the RP, encapsulated central a unicast packet, until

the allotment activity is completed

To appearance the complete process, Figure 17-15 shows an example. In the example, host H1 has

already abutting accumulation 228.8.8.8, as apparent in Figure 17-14. The afterward accomplish bout those

identified in Figure 17-15. Note that Footfall 3 represents the forwarding of the multicasts that were

encapsulated central Register letters at Footfall 2.

1. Host S1 sends multicasts to 228.8.8.8.

2. Router R1 encapsulates the multicasts, sending them central Register letters to the RP, R3.

3. R3, alive that it needs to advanced the multicast packets, de-encapsulates the packets and

sends them against H1. (This activity allows R1 and R3 to administer the multicasts while the

registration activity completes.) R5 assiduously the accumulation cartage to R4 and R4 assiduously it on

its LAN.

NOTE The characters (*,G) represents a distinct RPT. The * represents a wildcard, acceptation “any

source,” because the PIM-SM routers use this aggregate timberline behindhand of the antecedent of the packets.

For example, a packet beatific from any antecedent IP address, accession at the RP, and destined to group

228.8.8.8, would account the RP to use its (*,228.8.8.8) multicast acquisition table entries, because

these entries are allotment of the RPT for accumulation 228.8.8.8.

Sparse-Mode Acquisition Protocols 615

4. R3 joins the SPT for antecedent 10.1.1.10, accumulation 228.8.8.8, by sending a PIM-SM Accompany message

for accumulation (10.1.1.10,228.8.8.8) against the antecedent 10.1.1.10.

5. Aback R1 and R2 accept the PIM-SM Accompany bulletin from R2 requesting the accumulation cartage from

the source, they alpha forwarding accumulation cartage against the RP. At this point, R3 (the RP) now

receives this cartage on the SPT from the source. However, R1 is additionally still sending the Register

messages with encapsulated multicast packets to R3.

6. R3 sends unicast Register-Stop letters to R1. Aback R1 receives the Register-Stop

messages from R3, it stops sending the encapsulated unicast Register letters to R3.

Figure 17-15 Antecedent Allotment aback the RP Needs to Accept Packets Beatific to that Group

The activity may assume like a lot of trouble, but at the end of the process, multicasts are delivered

to the actual locations. The activity uses the able SPT from the antecedent to the RP, and the

shared timberline (*,228.8.8.8) from the RP to the subnets that charge to accept the traffic.

Note that the PIM protocols could accept aloof let a router abreast the source, such as R1 in this example,

continue to abbreviate multicasts central the unicast Register messages. However, it is inefficient

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

E0 .4

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

PIM-SM Join

(10.1.1.10, 228.8.8.8)

Multicast Packets (10.1.1.10, 228.8.8.8)

Unicast Register Messages

Unicast Register-Stop

4 10.1.10.3

2 Unicast Register

Message

Multicast Packet

(10.1.1.10, 228.8.8.8)

Multicast Cartage - RPT

(*, 228.8.8.8)

Multicast Traffic

(10.1.1.10, 228.8.8.8)

616 Affiliate 17: IP Multicast Routing

to accomplish R1 abbreviate every multicast packet, accomplish R3 de-encapsulate every packet, and then

make R3 advanced the traffic. So, PIM-SM has the RP, R3 in this case, accompany the group-specific tree

for that (S,G) combination.

Shared Administration Tree

In Figure 17-15, the accumulation cartage that flows over the aisle from the RP (R3) to R5 to R4 is called

a aggregate administration tree. It is additionally alleged a root-path timberline (RPT) because it is abiding at the RP. If

the arrangement has assorted sources for the aforementioned group, cartage from all the sources would aboriginal travel

to the RP (as apparent with the cartage from host S1 in Figure 17-14), and afresh biking bottomward this shared

RPT to all the receivers. Because all sources in the multicast accumulation use a accepted aggregate tree, a

wildcard characters of (*,G) is acclimated to analyze an RPT, area * represents all sources and G

represents the multicast accumulation address. The RPT for the accumulation 228.8.8.8 apparent in Figure 17-14

would be accounting as (*,228.8.8.8).

Example 17-7 shows the multicast avenue table access for R4 in Figure 17-15. On a Cisco router, the

show ip mroute command displays the multicast avenue table entries.

The estimation of the advice apparent in Archetype 17-7 is as follows:

■ The aboriginal band shows that the (*,G) access for the accumulation 228.8.8.8 was created 8 abnormal ago,

and if R4 does not advanced accumulation packets application this access in 2 account and 58 seconds, it will

expire. Every time R4 assiduously a packet, the timer is displace to 3 minutes. This access was created

because R4 accustomed an IGMP Accompany bulletin from H1.

■ The RP for this accumulation is 10.1.10.3 (R3). The S banderole indicates that this accumulation is application the

sparse-mode (PIM-SM) acquisition protocol. The C banderole indicates that R4 has a directly

connected accumulation affiliate for 228.8.8.8.

■ The admission interface for this (*,228.8.8.8) access is s0 and the RPF acquaintance is 10.1.6.5.

Note that for the SPT, the RPF interface is alleged based on the avenue to ability the RP, not the

route acclimated to ability a accurate source.

■ Accumulation cartage is forwarded out on the Ethernet0 interface. In this example, Ethernet0 was

added to the approachable interface account because an IGMP Report bulletin was accustomed on this

interface from H1. This interface has been in the forwarding accompaniment for 8 seconds. The Prune

timer indicates that if an IGMP Accompany is not accustomed afresh on this interface aural the next

2 account and 52 seconds, it will be removed from the approachable interface list.

Example 20-7 Multicast Avenue Table Access for the Accumulation 228.8.8.8 for R4

(*, 228.8.8.8), 00:00:08/00:02:58, RP 10.1.10.3, flags: SC

Incoming interface: Serial0, RPF nbr 10.1.6.5

Outgoing interface list:

Ethernet0, Forward/Sparse, 00:00:08/00:02:52

Sparse-Mode Acquisition Protocols 617

Steady-State Operation by Continuing to Accelerate Joins

To advance the forwarding accompaniment of interfaces, PIM-SM routers charge accelerate PIM Accompany messages

periodically. If a router fails to accelerate Joins periodically, PIM-SM moves interfaces aback to a pruned

state.

PIM-SM routers accept to advance the forwarding accompaniment on links based on two accepted criteria:

■ A afterwards router continues to accelerate PIM joins for the group.

■ A locally affiliated host still responds to IGMP Concern letters with IGMP Report

messages for the group.

Figure 17-16 shows an archetype in which R5 maintains the forwarding accompaniment of its articulation to R3 based

on both of these reasons. H2 has additionally abutting the aggregate timberline for 228.8.8.8. H1 had abutting earlier,

as apparent in Figures 17-14 and 17-15.

Figure 17-16 Host H2 Sends an IGMP Accompany Message

Example 17-8 shows the multicast avenue table access for R5 in Figure 17-16, with these two

interfaces in a forwarding state.

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

Periodic IGMP Reports

228.8.8.8

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

Periodic PIM Joins

(*, 228.8.8.8)

Periodic PIM Joins

(*, 228.8.8.8)

10.1.8.0/24

Periodic IGMP Reports

228.8.8.8

Multicast Traffic

(10.1.1.10, 228.8.8.8)

618 Affiliate 17: IP Multicast Routing

In Archetype 17-8, two interfaces are listed in the approachable interface list. The s1 interface is listed

because R5 has accustomed a PIM-SM Accompany bulletin from R4. In PIM-SM, the afterwards routers

need to accumulate sending PIM-SM Accompany letters every 60 abnormal to the upstream router. Aback R5

receives addition PIM-SM Accompany from R4 on its s1 interface, it resets the Clip timer to the default

value of 3 minutes. If R5 does not accept a PIM-SM Accompany from R4 afore R5’s Clip timer on that

interface expires, R5 places its s1 interface in a pruned accompaniment and stops forwarding the cartage on the

interface.

By contrast, R5’s e0 interface is listed as forwarding in R5’s approachable interface account because R5 has

received an IGMP Accompany bulletin from H2. Recall from Affiliate 16 that a multicast router sends an

IGMP accepted concern every 60 or 125 abnormal (depending on the IGMP version) on its LAN

interfaces. It charge accept at atomic one IGMP Report/Join bulletin as a acknowledgment for a group;

otherwise, it stops forwarding the accumulation cartage on the interface. Aback R5 receives addition IGMP

Report bulletin on its e0 interface, it resets the Clip timer for the access to the absence amount of

3 minutes.

Note additionally that on R5, the cancellation of the PIM Accompany from R4, or the IGMP Report on e0, triggers

R5’s charge to accelerate the PIM Accompany against the RP.

Examining the RP’s Multicast Acquisition Table

In the accepted accompaniment of the advancing example, as aftermost apparent in Figure 17-16, the RP (R3) has joined

the SPT for antecedent 10.1.1.10, accumulation 228.8.8.8. The RP additionally is the basis of the aggregate timberline for group

228.8.8.8. Archetype 17-9 shows both entries in R3’s multicast avenue table.

Example 17-8 Multicast Avenue Table Access for the Accumulation 228.8.8.8 for R5

(*,228.8.8.8), 00:00:05/00:02:59, RP 10.1.10.3, flags: SC

Incoming interface: Serial0, RPF nbr 10.1.5.3

Outgoing interface list:

Serial1, Forward/Sparse, 00:01:15/00:02:20

Ethernet0, Forward/Sparse, 00:00:05/00:02:55

Example 17-9 Multicast Avenue Table Access for the Accumulation 228.8.8.8 for R3

(*,228.8.8.8), 00:02:27/00:02:59, RP 10.1.10.3, flags: S

Incoming interface: Null, RPF nbr 0.0.0.0

Outgoing interface list:

Serial0, Forward/Sparse, 00:02:27/00:02:33

(10.1.1.10/32, 228.8.8.8), 00:02:27/00:02:33, flags: T

Incoming interface: Serial1, RPF nbr 10.1.3.2,

Outgoing interface list:

Outgoing interface list: Null

Sparse-Mode Acquisition Protocols 619

The aboriginal access shows the aggregate tree, as adumbrated by the S flag. Notice the admission interface is

Null because R3, as RP, is the basis of the tree. Also, the RPF acquaintance is listed as 0.0.0.0 for the

same reason. In added words, it shows that the shared-tree cartage for the accumulation 228.8.8.8 has

originated at this router and it does not depend on any added router for the shared-tree traffic.

The additional access shows the SPT access on R3 for multicast accumulation 228.8.8.8, antecedent 10.1.1.10. The

T banderole indicates that this access is for an SPT, and the antecedent is listed at the alpha of that same

line (10.1.1.10). The admission interface is s1 and the RPF acquaintance for the antecedent address

10.1.1.10 is 10.1.3.2.

As you can see, an RP uses the SPT to cull the cartage from the antecedent to itself and uses the shared

tree to advance the cartage bottomward to the PIM-SM routers that accept requested it.

Shortest-Path Timberline Switchover

PIM-SM routers could abide forwarding packets via the PIM-SM two-step process, whereby

sources accelerate packets to the RP, and the RP sends them to all added routers application the RPT. However,

one of the best alluring aspects of PIM-SM operations is that anniversary PIM-SM router can build

the SPT amid itself and the antecedent of a multicast accumulation and booty advantage of the best efficient

path accessible from the antecedent to the router. In Figure 17-16, R4 is accepting the accumulation cartage from

the antecedent via the aisle R1-R2-R3-R5-R4. However, it is accessible that it would be added efficient

for R4 to accept the accumulation cartage anon from R1 on R4’s s1 interface.

In the area “Completion of the Antecedent Allotment Process,” beforehand in this chapter, you saw

that the PIM-SM architecture allows an RP to body an SPT amid itself and the router that is directly

connected with the antecedent (also alleged the antecedent DR) to cull the accumulation traffic. Similarly, the PIMSM

design additionally allows any added PIM-SM router to body an SPT amid the router and the source

DR. This affection allows a PIM-SM router to abstain application the inefficient path, such as the one used

by R4 in Figure 17-16. Also, already the router starts accepting the accumulation cartage over the SPT, it can

send a Clip bulletin to the upstream router of the aggregate timberline to stop forwarding the cartage for

the group.

The catechism is, aback should a router about-face over from RPT to SPT? RFC 2362 for PIM-SM

specifies that, “The recommended action is to admit the about-face to the SP-tree afterwards accepting a

significant cardinal of abstracts packets during a defined time breach from a accurate source.” What

number should be advised as a cogent number? The RFC does not specify that. Cisco

routers, by default, about-face over from the RPT to the source-specific SPT afterwards they accept the first

packet from the aggregate tree.

NOTE You can change this behavior by configuring the all-around command ip pim sptthreshold

rate on any router for any group. Already the cartage amount exceeds the declared amount (in kbps),

the router joins the SPT. The command impacts the behavior alone on the router(s) on which it

is configured.

620 Affiliate 17: IP Multicast Routing

If a router is activity to about-face to SPT, why accompany the RPT first? In PIM-SM, a router does not know

the IP abode of a antecedent until it receives at atomic one packet for the accumulation from the source. After

it receives one packet on the RPT, it can apprentice the IP abode of a source, and initialize a switchover

to the SPT for that (source,group) combination.

With the absence Cisco PIM-SM operation, aback multicast packets activate accession on R4’s s0

interface via the aggregate tree, R4 attempts to about-face to the SPT for antecedent 10.1.1.10. Figure 17-17

shows the accepted steps.

Figure 17-17 R4 Initializing Switchover from RPT to SPT by Sending a PIM-SM Accompany to R1

The aboriginal three accomplish Figure 17-17 are as follows:

1. The antecedent (S1,10.1.1.10) sends a multicast packet to the first-hop router R1.

2. R1 assiduously the packet to the RP (R3).

3. The RP assiduously the packet to R4 via the aggregate tree.

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.5.0/24

10.1.4.0/24

10.1.7.0/24

10.1.6.0/24

E0 .4

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

10.1.8.0/24

E0 .5

Multicast Traffic

(10.1.1.10, 228.8.8.8)

Multicast Cartage - RPT

(*, 228.8.8.8)

Multicast Cartage - SPT

(10.1.1.10, 228.8.8.8)

5 Multicast Cartage -

(10.1.1.10, 228.8.8.8) PIM-SM Join

(10.1.1.10, 228.8.8.8)

Sparse-Mode Acquisition Protocols 621

At Footfall 3, R4 abstruse that the antecedent abode of the multicast accumulation 228.8.8.8 is 10.1.1.10. So,

besides forwarding the packet at Footfall 3, R4 can use that advice to accompany the SPT for group

228.8.8.8, from antecedent 10.1.1.10, application the afterward accomplish from Figure 17-17.

4. R4 consults its unicast acquisition table, finds the next-hop abode and approachable interface it

would use to ability antecedent 10.1.1.10, and sends the PIM-SM Accompany bulletin out that interface

(s1) to R1. This PIM-SM Accompany bulletin is accurately for the SPT of (10.1.1.10,228.8.8.8).

The Accompany campaign hop by hop until it alcove the antecedent DR.

5. As a aftereffect of the Join, R1 places its s1 interface in a forwarding accompaniment for SPT

(10.1.1.10,228.8.8.8). So, R1 starts forwarding multicasts from 10.1.1.10 to 228.8.8.8 out its

s1 interface as well.

R4 now has a multicast acquisition table access for the SPT, as apparent in Archetype 17-10.

In Archetype 17-10, you see two entries for the group. The J banderole (for join) on both the entries

indicates that the cartage was switched from RPT to SPT, and now the (S,G) access will be acclimated for

forwarding multicast packets for the group. Notice that the admission interfaces for the (*,G) entry

and (S,G) access are different.

Pruning from the Aggregate Tree

When a PIM-SM router has abutting a added able SPT, it may not charge to accept multicast

packets over the RPT any more. For example, aback R4 in Figure 17-17 notices that it is receiving

the accumulation cartage over RPT and SPT, it can and should ask the RP to stop sending the traffic.

To stop the RP from forwarding cartage to a afterwards router on the aggregate tree, the downstream

router sends a PIM-SM Clip bulletin to the RP. The Clip bulletin references the (S,G) SPT,

which identifies the IP abode of the source. Essentially, this clip agency the afterward to the RP:

Stop forwarding packets from the listed antecedent IP address, to the listed accumulation address,

down the RPT.

For example, in Figure 17-18, which continues the archetype apparent in Figure 17-17, R4 sends a

Prune out its s0 interface against R5. The Clip lists (S,G) access (10.1.1.10,228.8.8.8), and it sets

a bit alleged the RP-tree bit (RPT-bit). By ambience the RPT-bit in the Clip message, R4 informs

Example 17-10 Multicast Avenue Table Access for the Accumulation 228.8.8.8 for R4

(*,228.8.8.8), 00:02:36/00:02:57, RP 10.1.10.3, flags: SCJ

Incoming interface: Serial0, RPF nbr 10.1.6.5

Outgoing interface list:

Ethernet0, Forward/Sparse, 00:02:36/00:02:13

(10.1.1.10/32, 228.8.8.8), 00:00:23/00:02:33, flags: CJT

Incoming interface: Serial1, RPF nbr 10.1.4.1,

Outgoing interface list:

Ethernet0, Forward/Sparse, 00:00:23/00:02:37

622 Affiliate 17: IP Multicast Routing

R5 (the upstream router) that it has switched to SPT and the Clip bulletin is for the redundant

traffic for the accumulation 228.8.8.8, from 10.1.1.10, that R4 is accepting on the aggregate tree.

Figure 17-18 R4 Sends PIM-SM Clip with RP Bit Set to R5

To stop the packets from actuality beatific over the RPT to R4, R5 charge clip its interface s1 in the RPT

(*, 228.8.8.8). R5 may go on to accompany the SPT for (10.1.1.10,228.8.8.8.8) as well.

This concludes the advantage of the operations of PIM-SM. The abutting area covers some details

about how routers can apprentice the IP abode of the PIM RP.

Dynamically Finding RPs and Application Bombastic RPs

In a PIM-SM network, every router charge somehow apprentice the IP abode of an RP. A PIM-SM

router can use one of the afterward three methods to apprentice the IP abode of an RP:

■ The RP abode can be statically configured on all the PIM-SM routers with the Cisco IOS

global command ip pim rp-address address. This is the adjustment acclimated for the five-router

topology apparent in Figure 17-19.

■ The Cisco-proprietary Auto-RP agreement can be acclimated to baptize the RP and acquaint its IP

address so that all PIM-SM routers can apprentice its IP abode automatically.

■ A accepted BootStrap Router (BSR) agreement can be acclimated to baptize the RP and advertise

its IP abode so that all the PIM-SM routers can apprentice its IP abode automatically.

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

E0 .4

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

PIM-SM

RPT-bit Prune

(10.1.1.10, 228.8.8.8)

10.1.8.0/24

E0 .5

Multicast Traffic

(10.1.1.10, 228.8.8.8)

Sparse-Mode Acquisition Protocols 623

Additionally, because PIM-SM relies so heavily on the RP, it makes faculty to accept bombastic RPs.

Cisco IOS offers two methods of accouterment bombastic RPs, which are additionally covered in this section:

■ Anycast RP application the Multicast Antecedent Discovery Agreement (MSDP)

■ BootStrap Router (BSR)

Dynamically Finding the RP Application Auto-RP

Static RP agreement is suboptimal beneath the afterward conditions:

■ Aback an action has a ample cardinal of PIM-SM routers and the action wants to use

many altered RPs for altered groups, it becomes time arresting and bulky to

statically configure the IP addresses of abounding RPs for altered groups on all the routers.

■ Aback an RP fails or needs to be afflicted because a new RP is actuality installed, it becomes

extremely difficult in a statically configured PIM-SM area to about-face over to an alternative

RP after ample downtime.

Auto-RP provides an addition in which routers dynamically apprentice the unicast IP abode acclimated by

each RP. Auto-RP uses a two-step process, which is apparent in Figure 17-19 and Figure 17-20. In

the aboriginal step, the RP sends RP-Announce letters to the aloof multicast abode 224.0.1.39,

stating that the router is an RP. The RP-Announce bulletin additionally allows the router to acquaint the

multicast groups for which it is the RP, thereby acceptance some load-balancing of the RP workload

among altered routers. The RP continues to accelerate these RP-Announce letters every minute.

Figure 17-19 R3 Sends RP-Announce Messages

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

E0 .4

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

RP-Announce

(10.1.10.3, 224.0.1.39)

“Use my address

10.1.10.3 as an RP

address for all groups.”

10.1.8.0/24

E0 .5

Multicast Traffic

(10.1.1.10, 228.8.8.8)

624 Affiliate 17: IP Multicast Routing

For example, Figure 17-19 shows R3 as an RP that uses Auto-RP. R3 supports all multicast

groups in this case. The RP-Announce bulletin is apparent as Footfall 1, to articulation it with Footfall 2 in

Figure 17-20.

The additional footfall for Auto-RP requires that one router be configured as a mapping agent. The

mapping abettor is usually the aforementioned router that was alleged as an RP, but can be a altered PIMSM

router. The mapping abettor learns all the RPs and the multicast groups they anniversary support. Then,

the mapping abettor multicasts addition message, alleged RP-Discovery, that identifies the RP for

each ambit of multicast accumulation addresses. This bulletin goes to aloof multicast address

224.0.1.40. It is this RP-Discovery bulletin that absolutely informs the accepted router citizenry as

to which routers they should use as RPs.

For example, in Figure 17-20, R2 is configured as a mapping agent. To accept all RP-Announce

messages, R2 locally joins the acclaimed Cisco-RP-Announce multicast accumulation 224.0.1.39. In

other words, the mapping abettor has become a accumulation affiliate for 224.0.1.39 and is alert for

the accumulation traffic. Aback R2 receives the RP-Announce packets apparent in Figure 17-19, it examines

the packet, creates group-to-RP mappings, and maintains this advice in its cache, as shown

in Figure 17-20.

Figure 17-20 R2 Creates Group-to-RP Mappings and Sends Them in RP-Discovery Messages

R4 R5

R2 R3

.10

E0 .1

.1 .3

10.1.4.0/24 10.1.5.0/24

10.1.7.0/24

10.1.6.0/24

E0 .4

10.1.1.0/24

10.1.2.0/24 10.1.3.0/24

10.1.10.3

10.1.8.0/24

E0 .5

RP-Discovery 3

(10.1.2.2, 224.0.1.40)

Note these Group-To-RP

Mappings:

Group

224.0.0.0/4

RP Address

10.1.10.3

RP-Discovery

(10.1.3.2, 224.0.1.40)

Note these Group-To-RP

Mappings:

Group

224.0.0.0/4

RP Address

10.1.10.3

2 “I am a Mapping Agent.”

Group-To-RP Mappings:

Group

224.0.0.0/4

RP Address

10.1.10.3

Sparse-Mode Acquisition Protocols 625

At aboriginal glance, the charge for the mapping abettor may not be obvious. Why not aloof let the RPs

announce themselves to all the added routers? Well, if Auto-RP accurate alone one RP, or even

only one RP to abutment anniversary multicast group, the mapping abettor would be a decay of effort.

However, to abutment RP redundancy—in added words, to abutment assorted RPs that can act as RP

for the aforementioned multicast group—the Auto-RP mapping abettor decides which RP should be acclimated to

support anniversary accumulation at the moment. To do so, the mapping abettor selects the router with the highest

IP abode as an RP for the group. (Note that you can additionally configure assorted mapping agents, for

redundancy.)

As anon as Cisco routers are configured with PIM-SM and Auto-RP, they automatically accompany the

well-known Cisco-RP-Discovery multicast accumulation 224.0.1.40. That agency they are alert to the

group abode 224.0.1.40, and aback they accept a 224.0.1.40 packet, they apprentice group-to-RP

mapping advice and advance it in their cache. Aback a PIM-SM router receives an IGMP

Join bulletin for a accumulation or PIM-SM Accompany bulletin from a afterwards router, it checks the

group-to-RP mapping advice in its cache. Afresh it can advance as declared throughout

the PIM-SM explanations in this chapter, application that RP as the RP for that multicast group.

The afterward account summarizes the accomplish acclimated by Auto-RP:

1. Anniversary RP is configured to use Auto-RP and to advertise itself and its accurate multicast

groups via RP-Announce letters (224.0.1.39).

2. The Auto-RP mapping agent, which may or may not additionally be an RP router, gathers information

about all RPs by alert to the RP-Announce messages.

3. The mapping abettor builds a mapping table that lists the currently best RP for anniversary ambit of

multicast groups, with the mapping abettor acrimonious the RP with the accomplished IP abode if

multiple RPs abutment the aforementioned multicast groups.

4. The mapping abettor sends RP-Discover letters to 224.0.1.40 announcement the mappings.

5. All routers accept for packets beatific to 224.0.1.40 to apprentice the mapping advice and acquisition the

correct RP to use for anniversary multicast group.

Finally, one aftermost baby but important point deserves some absorption afore affective on to BSR.

Auto-RP creates a baby chicken-and-egg botheration in that the purpose of Auto-RP is to acquisition the

RPs, but to get the RP-Announce and RP-Discovery messages, PIM-SM routers would charge to

send a Accompany against the RP, which they do not apperceive yet. To affected this problem, Cisco added a

variation of PIM alleged sparse-dense mode. In PIM sparse-dense mode, a router uses PIM-DM

rules aback it does not apperceive the area of the RP, and PIM-SM rules aback it does apperceive the

location of the RP. So, beneath accustomed altitude with Auto-RP, the routers would use close mode

long abundant to apprentice the group-to-RP mappings from the mapping agent, and afresh about-face over to

sparse mode. Also, if any added multicast cartage occurred afore the routers abstruse of the RPs

using Auto-RP, the multicast packets would still be forwarded application dense-mode rules. (PIM

sparse-dense approach is configured per interface application the ip pim sparse-dense-mode interface

subcommand.)

Sources Sending Packets to the Rendezvous Point

Sources Sending Packets to the Rendezvous Point
PIM-SM uses a two-step process to initially deliver multicast packets from a particular source to
the hosts wanting to receive packets. Later, the process is improved beyond these initial steps. The
steps for the initial forwarding of multicasts with PIM-SM are as follows:
1. Sources send the packets to a router called the rendezvous point (RP).
2. The RP sends the multicast packets to all routers/hosts that have registered to receive packets
for that group. This process uses a shared tree.
NOTE In addition to these two initial steps, routers with local hosts that have sent an IGMP
Join for a group can go a step further, joining the source-specific tree for a particular (S,G) SPT.
This section describes the first of these two steps, in which the source sends packets to the RP. To
make that happen, the router connected to the same subnet as the source host must register with
the RP. The RP accepts the registration only if the RP knows of any routers or hosts that need to
receive a copy of those multicasts.
Figure 17-13 shows an example of the registration process in which the RP knows that no hosts
currently want the IP multicasts sent to group 228.8.8.8—no matter which source is sending them.
The configuration for this example is simple, with all the routers configured with the global
command ip multicast-routing and the interface command ip pim sparse-mode on all the
interfaces. Also, all routers have statically configured R3 as the RP by using the global command
ip pim rp-address 10.1.10.3. Usually, a loopback interface address is used as an RP address. The
loopback network 10.1.10.3/32 of R3 is advertised in the unicast routing protocol so that all the
routers know how to reach the RP.
Sparse-Mode Routing Protocols 611
Figure 17-13 Source Registration Process when RP Has Not Received a Request for the Group from Any
PIM-SM Router
The following three steps, referenced in Figure 17-13, describe the sequence of events for the
Source Registration process when the RP has not received a request for the group from any
PIM-SM router because no host has yet joined the group.
1. Host S1 begins sending multicasts to 228.8.8.8, and R1 receives those multicasts because it
connects to the same LAN.
2. R1 reacts by sending unicast PIM Register messages to the RP. The Register messages are
unicasts sent to the RP IP address, 10.1.10.3 in this case.
3. R3 sends unicast Register-Stop messages back to R1 because R3 knows that it does not have
any need to forward packets sent to 228.8.8.8.
In this example, the router near the source (R1) is attempting to register with the RP, but the RP
tells R1 not to bother any more, because no one wants those multicast messages. R1 has not
forwarded any of the native multicast messages at this point, in keeping with the PIM-SM strategy
of not forwarding multicasts until a host has asked for them. However, the PIM Register message
shown in Figure 17-13 encapsulates the first multicast packet. As will be seen in Figure 17-14, the
encapsulated packet would be forwarded by the RP had any senders been interested in receiving
the packets sent to that multicast group.
The source host may keep sending multicasts, so R1 needs to keep trying to register with the
RP in case some host finally asks to receive the packets. So, when R1 receives the Register-Stop
messages, it starts a 1-minute Register-Suppression timer. 5 seconds before the timer expires,
R1
R4 R5
R2 R3
.10
E0 .1
.1
S0
S1
.2
S0
.2
S1
.3
.5
.5
.1 .3
S1
S0
10.1.4.0/24 10.1.5.0/24
10.1.6.0/24
S0
S1
.4
.4
S0
S1
10.1.1.0/24
10.1.2.0/24 10.1.3.0/24
PIM Register
Unicast Register-Stop
RP
10.1.10.3
3
1
2
Unicast Register
Message
Multicast Packet
(10.1.1.10, 228.8.8.8)
Multicast Traffic
(10.1.1.10, 228.8.8.8)
S1
R1 sends another Register message with a flag set, called the Null-Register bit, without any
encapsulated multicast packets. As a result of this additional Register message, one of two things
will happen:
■ If the RP still knows of no hosts that want to receive these multicast packets, it sends another
Register-Stop message to R1, and R1 resets its Register-Suppression timer.
■ If the RP now knows of at least one router/host that needs to receive these multicast packets,
it does not reply to this briefer Register message. As a result, R1, when its timer expires, again
sends its multicast packets to R3 (RP) encapsulated in PIM Register messages.

Similarities Between PIM-DM and PIM-SM

Similarities Between PIM-DM and PIM-SM
PIM-SM has many similarities to PIM-DM. Like PIM-DM, PIM-SM uses the unicast routing table
to perform RPF checks—regardless of what unicast routing protocol populated the table. (Like
PIM-DM, the “protocol independent” part of the PIM acronym comes from the fact that PIM-SM
is not dependent on any particular unicast IP routing protocol.) In addition, PIM-SM also uses the
following mechanisms that are used by PIM-DM:
■ PIM Neighbor discovery through exchange of Hello messages.
■ Recalculation of the RPF interface when the unicast routing table changes.
■ Election of a DR on a multiaccess network. The DR performs all IGMP processes when
IGMPv1 is in use on the network.
■ The use of Prune Overrides on multiaccess networks.
■ Use of Assert messages to elect a designated forwarder on a multiaccess network. The winner
of the Assert process is responsible for forwarding unicasts onto that subnet.
These mechanisms are described in the “Operation of Protocol Independent Multicast Dense
Mode” section and thus are not repeated in this section.
NOTE The preceding list was derived, with permission, from Routing TCP/IP, Volume II, by
Jeff Doyle and Jennifer DeHaven Carroll.

Operation of Protocol Independent Multicast Sparse Mode

Operation of Protocol Independent Multicast Sparse Mode
PIM-SM works with a completely opposite strategy from that of PIM-DM, although the
mechanics of the protocol are not exactly opposite. PIM-SM assumes that no hosts want to receive
multicast packets until they specifically ask to receive them. As a result, until a host in a subnet
asks to receive multicasts for a particular group, multicasts are never delivered to that subnet. With
PIM-SM, downstream routers must request to receive multicasts using PIM Join messages. Also,
once they are receiving those messages, the downstream router must continually send Join messages
to the upstream router—otherwise, the upstream router stops forwarding, putting the link in a
pruned state. This process is opposite to that used by PIM-DM, in which the default is to flood
multicasts, with downstream routers needing to continually send Prunes or State Refresh messages
to keep a link in a pruned state.
PM-SM makes the most sense with a small percentage of subnets that need to receive packets sent
to any multicast group.

Sparse-Mode Routing Protocols

Sparse-Mode Routing Protocols
There are two sparse-mode routing protocols:
■ Protocol Independent Multicast Sparse Mode (PIM-SM)
■ Core-Based Tree (CBT)
This section covers the operation of PIM-SM.

Multicast Open Shortest Path First

Multicast Open Shortest Path First
MOSPF is defined in RFC 1584, “Multicast Extensions to OSPF,” which is an extension to the
OSPFv2 unicast routing protocol. The basic operation of MOSPF is described here:
■ MOSPF uses the group membership LSA, Type 6, which it floods throughout the originating
router’s area. As with unicast OSPF, all MOSPF routers in an area must have identical linkstate
databases so that every MOSPF router in an area can calculate the same SPT.
■ The SPT is calculated “on-demand,” when the first multicast packet for the group arrives.
■ Through the SPF calculation, all the routers know where the attached group members are,
based on the group membership LSAs.
■ After the SPF calculation is completed, entries are made into each router’s multicast
forwarding table.
■ Just like unicast OSPF, the SPT is loop free, and every router knows the upstream interface
and downstream interfaces. As a result, an RPF check is not required.
■ Obviously, MOSPF can only work with the OSPF unicast routing protocol. MOSPF is suitable
for small networks. As more hosts begin to source multicast traffic, routers have to perform
a higher number of Dijkstra algorithm computations, which demands an increasing level of
router CPU resources. Cisco IOS does not support MOSPF.

Distance Vector Multicast Routing Protocol

Distance Vector Multicast Routing Protocol
RFC 1075 describes Version 1 of DVMRP. DVMRP has many versions. The operation of DVMRP
is similar to PIM-DM. The major differences between PIM-DM and DVMRP are defined as
follows:
■ Cisco IOS does not support a full implementation of DVMRP; however, it does support
connectivity to a DVMRP network.
■ DVMRP uses its own distance vector routing protocol that is similar to RIPv2. It sends route
updates every 60 seconds and considers 32 hops as infinity. Use of its own routing protocol
adds more overhead to DVMRP operation compared to PIM-DM.
■ DVMRP uses Probe messages to find neighbors using the All DVMRP Routers group address
224.0.0.4.
■ DVMRP uses a truncated broadcast tree, which is similar to an SPT with some links
pruned.

LAN-Specific Issues with PIM-DM and PIM-SM

This area covers three baby capacity accompanying to operations that alone amount aback PIM is used

on LANs:

■ Clip Override

■ Assert messages

■ Appointed routers

Both PIM-DM and PIM-SM use these appearance in the aforementioned way.

Prune Override

In both PIM-DM and PIM-SM, the Clip action on multiaccess networks operates differently

from how it operates on point-to-point links. The acumen for this aberration is that aback one router

sends a Clip bulletin on a multiaccess network, added routers ability not appetite the articulation pruned by

the upstream router. Amount 17-11 shows an archetype of this problem, forth with the solution

through a PIM Join bulletin that is alleged a Clip Override. In this figure, R1 is forwarding the

group cartage for 239.9.9.9 on its fa0/0 interface, with R2 and R3 accepting the accumulation cartage on their

e0 interfaces. R2 does not accept any affiliated accumulation members, and its approachable interface list

would appearance null. The afterward account outlines the accomplish in argumentation apparent in Amount 17-11, in which R3

needs to accelerate a Clip Override:

1. R2 sends a Clip for accumulation 239.9.9.9 because R2 has a absent approachable interface account for the group.

2. R1, acumen that it accustomed the Clip on a multiaccess network, knows that added routers

might still appetite to get the messages. So, instead of anon pruning the interface, R1 sets

a 3-second timer that charge expire afore R1 will clip the interface.

Dense-Mode Acquisition Protocols 605

3. R3 additionally receives the Clip bulletin beatific by R2, because Clip letters are multicast to All-

PIM-Routers accumulation abode 224.0.0.13. R3 still needs to get cartage for 239.9.9.9, so R3 sends

a Join bulletin on its e0 interface.

4. (Not apparent in Amount 17-11) R1 receives the Join bulletin from R3 afore removing its LAN

interface from the approachable interface list. As a result, R1 does not clip its Fa0/0 interface.

Figure 17-11 Clip Override

This action is alleged Clip Override because R3 overrides the Clip beatific by R2. The Prune

Override is absolutely a Join message, beatific by R3 in this case. The bulletin itself is no altered from

a accustomed Join. As continued as R1 receives a Join bulletin from R3 afore its 3-second timer expires,

R3 continues to accept cartage after interruption.

Assert Message

The final PIM-DM bulletin covered in this affiliate is the PIM Assert message. The Assert

message is acclimated to anticipate ashen accomplishment aback added than one router attaches to the aforementioned LAN.

Rather than sending assorted copies of anniversary multicast packet assimilate the LAN, the PIM Assert

message allows the routers to negotiate. The champ gets the appropriate to be amenable for forwarding

multicasts assimilate the LAN.

Figure 17-12 shows an archetype of the charge for the Assert message. R2 and R3 both attach to the

same LAN, with H1 actuality an alive affiliate of the accumulation 227.7.7.7. Both R2 and R3 are receiving

the accumulation cartage for 227.7.7.7 from the antecedent 10.1.1.10.

R2 R3

E0 E0

I got a clip on a

multiaccess net – set

my 3 additional timer!

I accept a absent OIL for

239.9.9.9 – Prune!

I’d bigger send

another Join before

R1 prunes 239.9.9.9

Prune

239.9.9.9

Join, 239.9.9.9

(Prune Override)

Member,

239.9.9.9

Fa0/0

Fa0/1

606 Affiliate 17: IP Multicast Routing

Figure 17-12 R2 and R3 Sending Assert Messages

The ambition of the Assert bulletin is to accredit the albatross of forwarding accumulation cartage on the

LAN to the router that is abutting to the source. Aback R2 and R3 accept accumulation cartage from the

source on their s0 interfaces, they advanced it on their e0 interfaces. Both of them accept their s0

interfaces in the admission interface account and e0 interfaces in the approachable interface list. Now, R2

and R3 accept a multicast packet for the accumulation on their e0 interfaces, which will account them to

send an Assert bulletin to boldness who should be the forwarder.

The Assert action picks a champ based on the acquisition agreement and metric acclimated to acquisition the route

to ability the unicast abode of the source. In this example, that agency that R2 or R3 will win based

on the routes they anniversary use to ability 10.1.1.10. R2 and R3 accelerate and accept Assert letters that

include their corresponding authoritative distances of the acquisition protocols acclimated to apprentice the route

that matches 10.1.1.10, as able-bodied as the metric for those routes. The routers on the LAN compare

their own acquisition agreement authoritative ambit and metrics to those abstruse in the Assert

messages. The champ of the Assert action is bent as follows:

1. The router announcement the everyman authoritative ambit of the acquisition agreement acclimated to learn

the avenue wins.

2. If a tie, the router with the everyman advertised acquisition agreement metric for that avenue wins.

3. If a tie, the router with the accomplished IP abode on that LAN wins.

Designated Router

PIM Hello letters are additionally acclimated to accept a appointed router (DR) on a multiaccess network. A

PIM-DM or PIM-SM router with the accomplished IP abode becomes a DR.

R2 R3

.2 E0 .3 E0

.1 .1

S0 S1

Assert Assert

Member of Group

227.7.7.7

10.1.1.0/24

10.1.4.0/24

10.1.2.0/24 10.1.3.0/24

Multicast Traffic

Destination Address: 227.7.7.7

Source Address: 10.1.1.10

.10

Dense-Mode Acquisition Protocols 607

The PIM DR abstraction applies mainly aback IGMPv1 is used. IGMPv1 does not accept a mechanism

to accept a Querier—that is to say that IGMPv1 has no way to adjudge which of the abounding routers on

a LAN should accelerate IGMP Queries. Aback IGMPv1 is used, the PIM DR is acclimated as the IGMP

Querier. IGMPv2 can anon accept a Querier (the router with the everyman IP address), so the PIM

DR is not acclimated as the IGMP Querier aback IGMPv2 is used.

Note that on a LAN, one router ability win the Assert action for a accurate (S,G) SPT, while

another ability become the IGMP Querier (PIM DR for IGMPv1, IGMP Querier for IGMPv2). The

winner of the Assert action is amenable for forwarding multicasts assimilate the LAN, admitting the

IGMP Querier is amenable for managing the IGMP action by actuality amenable for sending

IGMP Query letters on the LAN. Note additionally that the IGMPv2 Querier acclamation chooses the

lowest IP address, and the Assert action uses the accomplished IP abode as a tiebreaker, authoritative it

slightly added acceptable that altered routers are called for anniversary function.

Summary of PIM-DM Messages

This area concludes the advantage of PIM-DM. Table 17-2 lists the key PIM-DM messages

covered in this chapter, forth with a abrupt analogue of their use.

Table 17-2 Summary of PIM-DM Messages

PIM Bulletin Definition

Hello Acclimated to anatomy acquaintance adjacencies with added PIM routers, and to

maintain adjacencies by ecology for accustomed Hellos from each

neighbor. Additionally acclimated to accept a PIM DR on multiaccess networks.

Prune Acclimated to ask a adjoining router to abolish the articulation over which the

Prune flows from that adjoining router’s approachable interface account for a

particular (S,G) SPT.

State Refresh Acclimated by a after router, beatific to an upstream router on an RPF

interface, to account the upstream router to displace its Clip timer. This

allows the after router to advance the pruned accompaniment of a link, for

a accurate (S,G) SPT.

Assert Acclimated on multiaccess networks to actuate which router wins the right

to advanced multicasts assimilate the LAN, for a accurate (S,G) SPT.

Prune Override (Join) On a LAN, a router may multicast a Clip bulletin to its upstream

routers. Added routers on the aforementioned LAN, absent to anticipate the

upstream router from pruning the LAN, anon accelerate addition Join

message for the (S,G) SPT. (The Clip Override is not absolutely a Prune

Override message—it is a Join. This is the alone purpose of a Join

message in PIM-DM, per RFC 3973.)

Graft/Graft-Ack Aback a pruned articulation needs to be added aback to an (S,G) SPT, a router

sends a Graft bulletin to its RPF neighbor. The RPF neighbor

acknowledges with a Graft-Ack.

Dense-Mode Acquisition Protocols

There are three dense-mode acquisition protocols:

■ Agreement Absolute Multicast Close Mode (PIM-DM)

Dense-Mode Acquisition Protocols 593

■ Distance Vector Multicast Acquisition Agreement (DVMRP)

■ Multicast Open Shortest Aisle Aboriginal (MOSPF)

This area covers the operation of PIM-DM in detail and provides an overview of DVMRP and

MOSPF.

Operation of Agreement Absolute Multicast Close Mode

Protocol Absolute Multicast (PIM) defines a alternation of agreement letters and rules by which

routers can accommodate able forwarding of multicast IP packets. PIM ahead existed as a

Cisco-proprietary protocol, although it has been offered as an alpha agreement via RFCs

2362, 3446, and 3973. The PIM blueprint spell out the rules mentioned in the beforehand examples

in this chapter—things like the RPF check, the PIM dense-mode argumentation of calamity multicasts until

routers accelerate Clip messages, and the PIM Sparse-mode argumentation of not forwarding multicasts

anywhere until a router sends a Accompany message. This area describes the PIM-DM protocols in

more detail.

PIM gets its name from its adeptness to use the unicast IP acquisition table for its RPF check—

independent of whatever unicast IP acquisition protocol(s) was acclimated to body the unicast acquisition table

entries. In fact, the name “PIM” absolutely says as abundant about the two added dense-mode protocols—

DVMRP and MOSPF—as it does about PIM. These added two protocols do not use the unicast IP

routing table for their RPF checks, instead architecture their own absolute tables. PIM simply

relies on the unicast IP acquisition table, absolute of which unicast IP acquisition agreement congenital a

particular access in the acquisition table.

Forming PIM Adjacencies Application PIM Accost Messages

PIM routers anatomy adjacencies with adjoining PIM routers for the aforementioned accepted reasons, and

with the aforementioned accepted mechanisms, as abounding added acquisition protocols. PIMv2, the accepted version

of PIM, sends Accost letters every 30 abnormal (default) on every interface on which PIM is

configured. By accepting Hellos on the aforementioned interface, routers ascertain neighbors, establish

adjacency, and advance adjacency. PIMv2 Hellos use IP agreement cardinal 103 and reserved

multicast destination abode 224.0.0.13, alleged the All-PIM-Routers multicast address. The Hello

messages accommodate a Holdtime value, about three times the sender’s PIM accost interval. If the

receiver does not accept a Accost bulletin from the sender during the Holdtime period, it considers

the sending acquaintance to be dead.

NOTE The beforehand version, PIMv1, does not use Hellos, instead application a PIM Concern message.

PIMv1 letters are encapsulated in IP packets with agreement cardinal 2 and use the multicast

destination abode 224.0.0.2.

594 Affiliate 17: IP Multicast Routing

As you will see in the afterward sections, establishing and advancement adjacencies with directly

connected neighbors is actual important for the operation of PIM. A PIM router sends added PIM

messages alone on interfaces on which it has accepted alive PIM neighbors.

Source-Based Administration Trees

Dense-mode acquisition protocols are acceptable for close cartography in which there are abounding multicast

group associates about to the absolute cardinal of hosts in a network. Aback a PIM-DM router receives

a multicast packet, it aboriginal performs the RPF check. If the RPF analysis succeeds, the router forwards

a archetype of the packet to all the PIM neighbors except the one on which it accustomed the packet. Each

PIM-DM router repeats the action and floods the absolute arrangement with the accumulation traffic. Ultimately,

the packets are abounding to all blade routers that accept no afterwards PIM neighbors.

The argumentation declared in the antecedent branch absolutely describes the concepts abaft what PIM

calls a source-based administration tree. It is additionally sometimes alleged a shortest-path timberline (SPT), or

simply a antecedent tree. The timberline defines a aisle amid the antecedent host that originates the multicast

packets and all subnets that charge to accept a archetype of the multicasts beatific by that host. The timberline uses

the antecedent as the root, the routers as the nodes in the tree, and the subnets affiliated to the routers

as the branches and leaves of the tree. Figure 17-3, beforehand in the chapter, shows the abstraction behind

an SPT.

The agreement appropriate on the three routers in Figure 17-3 is easy—just add the all-around command

ip multicast-routing on anniversary router and the interface command ip pim dense-mode on all the

interfaces of all the routers.

PIM-DM ability accept a altered source-based administration timberline for anniversary aggregate of antecedent and

multicast group, because the SPT will alter based on the area of the antecedent and the locations

of the hosts alert for anniversary multicast accumulation address. The characters (S,G) refers to a particular

SPT, or to an alone router’s allotment of a accurate SPT, area S is the source’s IP abode and G

is the multicast accumulation address. For example, the (S,G) characters for the archetype in Figure 17-3

would be accounting as (10.1.1.10, 226.1.1.1).

Example 17-1 shows allotment of the (S,G) SPT access on R3, from Figure 17-3, for the (10.1.1.0,

226.1.1.1) SPT. Host S1 is sending packets to 226.1.1.1, and host H2 sends an IGMP Join

message for the accumulation 226.1.1.1. Archetype 17-1 shows a allotment of R3’s multicast acquisition table, as

displayed application the appearance ip mroute command.

Example 17-1 Multicast Avenue Table Access for the Accumulation 226.1.1.1 for R3

(10.1.1.10/32, 226.1.1.1), 00:00:12/00:02:48, flags: CT

Incoming interface: Serial0/1, RPF nbr 10.1.4.1

Outgoing interface list:

FastEthernet0/0, Forward/Dense, 00:00:12/00:00:00

Dense-Mode Acquisition Protocols 595

The estimation of the advice apparent in Archetype 17-1 is as follows:

■ The aboriginal band shows that the (S, G) access for (10.1.1.10/32, 226.1.1.1) has been up for

12 seconds, and that if R3 does not advanced an (S, G) packet in 2 account and 48 seconds, it

will expire. Every time R3 assiduously a packet application this entry, the timer is displace to 3 minutes.

■ The C banderole indicates that R3 has a anon affiliated accumulation affiliate for 226.1.1.1. The T flag

indicates that the (S,G) cartage is forwarded on the shortest-path tree.

■ The admission interface for the accumulation 226.1.1.1 is s0/1 and the RPF acquaintance (the next-hop IP

address to go in the about-face administration against the antecedent abode 10.1.1.10) is 10.1.4.1.

■ The accumulation cartage is forwarded out on the fa0/0 interface. This interface has been in the

forwarding accompaniment for 12 seconds. The additional timer is listed as 00:00:00, because it cannot

expire with PIM-DM, as this interface will abide to advanced cartage until pruned.

The abutting two sections appearance how PIM-DM routers use advice abstruse from IGMP to

dynamically aggrandize and arrangement the source-based administration copse to amuse the needs of the

group users.

Prune Message

PIM-DM creates a new SPT aback a antecedent aboriginal sends multicast packets to a new multicast group

address. The SPT includes all interfaces except RPF interfaces, because PIM-DM assumes that all

hosts charge to accept a archetype of anniversary multicast packet. However, some subnets may not charge a copy

of the multicasts, so PIM-DM defines a action by which routers can abolish interfaces from an

SPT by application PIM Clip messages.

For example, in Figure 17-3, hosts H1 and H2 charge a archetype of the multicast packets beatific to

226.1.1.1. However, as shown, aback R2 gets the multicast from R1, R2 afresh assiduously the

multicasts to R3. As it turns out, R3 is bottomward the packets for the accumulation cartage from 10.1.1.1,

sent to 226.1.1.1, because those packets abort R3’s RPF check. In this case, R3 can account R2 to

remove its s0/1 interface from its approachable interface account for (10.1.1.10, 226.1.1.1) by sending a

NOTE The multicast acquisition table flags mentioned in this list, as able-bodied as others, are

summarized in Table 17-6 in the “Foundation Summary” area of this chapter.

NOTE According to PIM-DM specifications, multicast avenue tables alone charge (S,G) entries.

However, for anniversary (S,G) entry, a Cisco router creates a (*,G) access as a ancestor entry, for design

efficiency. The (*,G) access is not acclimated for forwarding the multicast cartage for a accumulation that uses

PIM-DM. Therefore, for artlessness and clarity, the (*,G) entries are not apparent in the examples

that use PIM-DM. Had you congenital the aforementioned arrangement as illustrated in Figure 17-3, and configured

PIM-DM, the (*,G) entries would additionally be listed in the appearance ip mroute command output.

596 Affiliate 17: IP Multicast Routing

Prune bulletin to R2. As a result, R2 will not advanced the multicasts to R3, thereby abbreviation the

amount of ashen bandwidth.

The afterward is a added academic analogue of a PIM Clip message:

The PIM Clip bulletin is beatific by one router to a additional router to account the second

router to abolish the articulation on which the Clip is accustomed from a accurate (S,G) SPT.

Figure 17-6 shows the aforementioned internetwork and archetype as Figure 17-3, but with R3’s Prune

messages beatific to R2.

Figure 17-6 R3 Sends a Clip Bulletin to R2

As a aftereffect of the Clip bulletin from R3 to R2, R2 will clip its s0/1 interface from the SPT for

(10.1.1.10,226.1.1.1). Archetype 17-2 shows the multicast avenue table access for R2 in Figure 17-6,

with the band that shows the pruned accompaniment highlighted.

NOTE The appellation approachable interface account refers to the account of interfaces in a forwarding state,

listed for an access in a router’s multicast acquisition table.

Fa0/0

10.1.1.0/24

10.1.2.0/24

10.1.5.0/24

10.1.6.0/24

10.1.3.0/24

10.1.4.0/24

S0/0 S0/0 .2

S0/1

S0/0

S0/1

Packet: Antecedent = 10.1.1.0,

Destination = 226.1.1.1

R1 R2

.1 .2

.10

Prune

My S0/1 interface

was pruned – do not

forward the packets

to R3 any more!

Dense-Mode Acquisition Protocols 597

Most of the advice apparent in Archetype 17-2 is agnate to the advice apparent in

Example 17-1. Apprehension the Serial0/1 advice apparent beneath the approachable interface list. It

shows that this interface was pruned 8 abnormal ago because R3 beatific a Clip bulletin to R2.

This agency that, at this time, R2 is not forwarding cartage for 226.1.1.1 on its s0/1 interface.

Because PIM-DM’s inherent addiction is to flood cartage through an internetwork, the pruned s0/1

interface listed in Archetype 17-2 will be afflicted aback to a forwarding accompaniment afterwards 2 account and

52 seconds. In PIM-DM, aback a router receives a Clip bulletin on an interface, it starts a

(default) 3-minute Clip timer, counting bottomward to 0. Aback the Clip timer expires, the router

changes the interface to a forwarding accompaniment again. If the afterwards router does not appetite the

traffic, it can afresh accelerate a Clip message. This affection keeps a afterwards router acquainted that the

group cartage is accessible on a accurate interface from the upstream neighbor.

Note that a multicast router can accept added than one interface in the approachable interface list, but it

can accept alone one interface in the admission interface list. The alone interface in which a router will

receive and action multicasts from a accurate antecedent is the RPF interface. Routers still perform

an RPF check, with the admission interface advice in the alpha of the appearance ip mroute

output advertence the RPF interface and neighbor.

PIM-DM: Reacting to a Bootless Link

When links fail, or any added changes affect the unicast IP acquisition table, PIM-DM needs to update

the RPF interfaces based on the new unicast IP acquisition table. Because the RPF interface may

change, (S,G) entries may additionally charge to account altered interfaces in the approachable interface list. This

section describes an archetype of how PIM-DM reacts.

Figure 17-7 shows an archetype in which the articulation amid R1 and R3, originally illustrated in

Figure 17-6, has failed. Afterwards the unicast acquisition agreement converges, R3 needs to amend its RPF

neighbor IP abode from 10.1.4.1 (R1) to 10.1.3.2 (R2). Additionally in this case, H1 has issued an IGMP

Leave message.

Example 17-2 Multicast Avenue Table Access for the Accumulation 226.1.1.1 for R2

(10.1.1.10/32, 226.1.1.1), 00:00:14/00:02:46, flags: CT

Incoming interface: Serial0/0, RPF nbr 10.1.2.1

Outgoing interface list:

FastEthernet0/0, Forward/Dense, 00:00:14/00:00:00

Serial0/1, Prune/Dense, 00:00:08/00:02:52

NOTE PIMv2 offers a bigger band-aid to advancement the pruned accompaniment of an interface, using

State Brace messages. These letters are covered afterwards in the chapter, in the area “Steady-

State Operation and the Accompaniment Brace Message.”

598 Affiliate 17: IP Multicast Routing

Figure 17-7 Direct Articulation Amid R1 and R3 Is Bottomward and Host H1 Sends an IGMP Leave Message

Example 17-3 shows the consistent multicast avenue table access for R3 in Figure 17-7. Note that the

RPF interface and acquaintance IP abode has afflicted to point to R2.

Example 17-3 shows how R3’s appearance of the (10.1.1.10,226.1.1.1) SPT has changed. However, R2

had pruned its s0/1 interface from that SPT, as apparent in Figure 17-6. So, R2 needs to change its

s0/1 interface aback to a forwarding accompaniment for SPT (10.1.1.10, 226.1.1.1). Archetype 17-4 shows the

resulting multicast avenue table access for (10.1.1.10, 226.1.1.1) in R2.

Example 17-3 Multicast Avenue Table Access for the Accumulation 226.1.1.1 for R3

(10.1.1.10/32, 226.1.1.1), 00:02:16/00:01:36, flags: CT

Incoming interface: Serial0/0, RPF nbr 10.1.3.2

Outgoing interface list:

FastEthernet0/0, Forward/Dense, 00:02:16/00:00:00

Example 17-4 Multicast Avenue Table Access for the Accumulation 226.1.1.1 for R2

(10.1.1.10/32, 226.1.1.1), 00:03:14/00:02:38, flags: T

Incoming interface: Serial0/0, RPF nbr 10.1.2.1

Outgoing interface list:

Serial0/1, Forward/Dense, 00:02:28/00:00:00

Fa0/0

10.1.1.0/24

10.1.2.0/24

10.1.5.0/24

10.1.6.0/24

10.1.3.0/24

IGMP Leave

for 226.1.1.1

S0/0 S0/0 .2

S0/1

S0/0

R1 R2

.1 .2

.10

Destination Address: 226.1.1.1

Source Address: 10.1.1.10

Multicast Traffic

Dense-Mode Acquisition Protocols 599

In Archetype 17-4, apprehension the approachable interface account for R2. R2 has now removed interface fa0/0

from the approachable interface account and chock-full forwarding cartage on the interface because it received

no acknowledgment to the IGMP Group-Specific concern for accumulation 226.1.1.1. As a result, R2 has also

removed the C banderole (C acceptation “connected”) from its multicast acquisition table access for (10.1.1.10,

226.1.1.1). Additionally, R2 assiduously the cartage on its s0/1 interface against R3 because R3 is still

forwarding cartage on its fa0/0 interface and has not yet beatific a Clip bulletin to R2.

Rules for Pruning

This area explains two key rules that a PIM-DM router charge chase to adjudge aback it can

request a prune. Afore answer addition archetype of how PIM-DM reacts to changes in an

internetwork, a brace of new multicast agreement charge be defined. To abridge the wording, the

following statements ascertain upstream router and afterwards router from the angle of a

router alleged R1.

■ R1’s upstream router is the router from which R1 receives multicast packets for a

particular SPT.

■ R1’s afterwards router is a router to which R1 assiduously some multicast packets for a

particular SPT.

For example, R1 is R2’s upstream router for the packets that S1 is sending to 226.1.1.1 in Figure 17-7.

R3 is R2’s afterwards router for those aforementioned packets, because R2 sends those packets to R3.

PIM-DM routers can accept to accelerate a Clip bulletin for abounding reasons, one of which was covered

earlier with attention to Figure 17-6. The capital affidavit are abbreviated here:

■ Aback accepting packets on a non-RPF interface.

■ Aback a router realizes that both of the afterward are true:

— No locally affiliated hosts in a accurate accumulation are alert for packets.

— No afterwards routers are alert for the group.

This area shows the argumentation abaft the additional acumen for sending prunes. At this point in

the account of Figures 17-6 and 17-7, the alone host that needs to accept packets beatific to

226.1.1.1 is H2. What would the PIM-DM routers in this arrangement do if H2 leaves group

226.1.1.1? Figure 17-8 shows aloof such an example, with H2 sending an IGMP Leave message

for accumulation 226.1.1.1. Figure 17-8 shows how PIM-DM uses this advice to dynamically

update the SPT.

NOTE R2 afflicted its s0/1 to a forwarding accompaniment because of a PIM Affix bulletin beatific by R3.

The accessible area “Graft Message” explains the details.

600 Affiliate 17: IP Multicast Routing

Figure 17-8 R3 and R2 Sending Clip Messages

Figure 17-8 shows three steps, with the argumentation in Accomplish 2 and 3 actuality agnate but actual important:

1. H2 leaves the multicast accumulation by application an IGMP Leave message.

2. R3 uses an IGMP Concern to affirm that no added hosts on the LAN appetite to accept cartage for

group 226.1.1.1. So, R3 sends a Prune, referencing the (10.1.1.20, 226.1.1.1) SPT, out its RPF

interface R2.

3. R2 does not accept any locally affiliated hosts alert for accumulation 226.1.1.1. Now, its only

downstream router has beatific a Clip for the SPT with antecedent 10.1.1.10, accumulation 226.1.1.1.

Therefore, R2 has no acumen to charge packets beatific to 226.1.1.1 any more. So, R2 sends a Prune,

referencing the (10.1.1.20, 226.1.1.1) SPT, out its RPF interface R1.

After the pruning is complete, both R3 and R2 will not be forwarding cartage beatific to 226.1.1.1 from

source 10.1.1.10. In the routers, the appearance ip mroute command shows that actuality application the P (prune)

flag, which agency that the router has absolutely pruned itself from that accurate (S,G) SPT.

Example 17-5 shows R3’s command achievement with a absent approachable interface list.

Example 17-5 Multicast Avenue Table Access for the Accumulation 226.1.1.1 for R3

(10.1.1.10/32, 226.1.1.1), 00:03:16/00:01:36, flags: PT

Incoming interface: Serial0/0, RPF nbr 10.1.3.2

Outgoing interface list: Null

Fa0/0

10.1.1.0/24

10.1.2.0/24

10.1.5.0/24

10.1.6.0/24

10.1.3.0/24

IGMP Leave

for 226.1.1.1

S0/0 S0/0 .2

S0/1

S0/0

R1 R2

.1 .2

.10

Destination Address: 226.1.1.1

Source Address: 10.1.1.10

Prune

Multicast Traffic

Dense-Mode Acquisition Protocols 601

After all the accomplish in Figure 17-8 accept been completed, R1 additionally does not charge to accelerate packets sent

by 10.1.1.10 to 226.1.1.1 out any interfaces. Afterwards accepting a Clip bulletin from R2, R1 has also

updated its approachable interface list, which shows that there is alone one approachable interface and that

it is in the pruned accompaniment at this time. Archetype 17-6 shows the details.

Of accurate absorption in the output, R1 has additionally set the C flag, but for R1 the C banderole does not indicate

that it has anon affiliated accumulation members. In this case, the aggregate of a C banderole and an RPF

neighbor of 0.0.0.0 indicates that the affiliated accessory is the antecedent for the group.

In reality, there is no abstracted Clip bulletin and Accompany message; instead, PIM-DM and PIM-SM

use a distinct bulletin alleged a Join/Prune message. A Clip bulletin is absolutely a Join/Prune

message with a accumulation abode listed in the Clip field, and a Accompany bulletin is a Join/Prune message

with a accumulation abode listed in the Accompany field.

Steady-State Operation and the Accompaniment Brace Message

As mentioned briefly beforehand in the chapter, with PIM-DM, an interface stays pruned alone for

3 account by default. Clip letters account a accurate antecedent and accumulation (in added words, a particular

(S,G) SPT). Whenever a router receives a Clip message, it finds the analogous (S,G) SPT entry

and marks the interface on which the Clip bulletin was accustomed as “pruned.” However, it also

sets a Clip timer, absence 3 minutes, so that afterwards 3 minutes, the interface is placed into a forwarding

state again.

So, what happens with PIM-DM and pruned links? Well, the all-important links are pruned,

and 3 account afterwards they are added back. Added multicasts flow, and the links are pruned. Then

they are added back. And so on. So, aback Cisco created PIM V2 (published as experimental

RFC 3973), it included a affection alleged accompaniment refresh. Accompaniment Brace letters can anticipate this

rather inefficient behavior in PIM-DM adaptation 1 of pruning and automatically unpruning

interfaces.

Figure 17-9 shows an archetype that begins with the aforementioned accompaniment as the arrangement declared at the end

of the above-mentioned section, “Rules for Pruning,” area the articulation amid R1 and R2 and the link

between R2 and R3 accept been pruned. Almost 3 account accept passed, and the links are about to

be added to the SPT afresh due to the cessation of the Clip timers.

Example 17-6 Multicast Avenue Table Access for the Accumulation 226.1.1.1 for R1

(10.1.1.10/32, 226.1.1.1), 00:08:35/00:02:42, flags: CT

Incoming interface: FastEthernet0/0, RPF nbr 0.0.0.0

Outgoing interface list:

Serial0/0, Prune/Dense, 00:00:12/00:02:48

602 Affiliate 17: IP Multicast Routing

Figure 17-9 How PIM-DM Adaptation 2 Uses Accompaniment Brace Messages

The PM Accompaniment Brace bulletin can be sent, aloof afore a neighbor’s Clip timer expires, to keep

the interface in a pruned state. In Figure 17-9, the afterward accomplish do aloof that:

1. R3 monitors the time aback it beatific the aftermost Clip to R2. Aloof afore the Clip timer expires,

R3 decides to accelerate a Accompaniment Brace bulletin to R2.

2. R3 sends the Accompaniment Brace bulletin to R2, referencing SPT (10.1.1.10, 226.1.1.1).

3. R2 reacts by resetting its Clip timer for the interface on which it accustomed the Accompaniment Refresh

message.

4. Because R2 had additionally pruned itself by sending a Clip bulletin to R1, R2 additionally uses State

Refresh letters to acquaint R1 to leave its s0/0 interface in a pruned state.

As continued as R3 keeps sending a Accompaniment Brace bulletin afore the Clip timer on the upstream

router (R2) expires, the SPT will abide stable, and there will not be the alternate times of flooding

of added multicasts for that (S,G) tree.

Graft Message

When new hosts accompany a group, routers may charge to change the accepted SPT for a accurate (S,G)

entry. With PIM-DM, one advantage could be to delay on the pruned links to expire. For example, in

Figure 17-9, R3 could artlessly abdicate sending Accompaniment Brace messages, and aural 3 account at most,

Fa0/0

10.1.1.0/24

10.1.2.0/24

10.1.5.0/24

10.1.6.0/24

10.1.3.0/24

S0/0 S0/0 .2

S0/1

S0/0

R1 R2

.1 .2

.10

Destination Address: 226.1.1.1

Source Address: 10.1.1.10

I got a accompaniment refresh

message – displace my

prune timer on S0/1!

My Clip timer for (10.1.1.10,

226.1.1.1), interface S0/0, is

about to expire. Accelerate a staterefresh

message!

State

Refresh

State

Refresh

Multicast Traffic

Dense-Mode Acquisition Protocols 603

R3 would be accepting the multicast packets for some (S,G) SPT again. However, cat-and-mouse on the

(default) 3-minute Clip timer to expire is not actual efficient. To acquiesce routers to “unprune” a

previously pruned interface from an SPT, PIM-DM includes the Affix message, which is defined

as follows:

A router sends a Affix bulletin to an upstream neighbor—a acquaintance to which it had

formerly beatific a Clip message—causing the upstream router to put the articulation aback into a

forwarding accompaniment (for a accurate (S,G) SPT).

Figure 17-10 shows an archetype that uses the aforementioned advancing archetype network. The action shown

in Figure 17-10 begins in the aforementioned accompaniment as declared at the end of the above-mentioned section, “Steady-

State Operation and the Accompaniment Brace Message.” Neither host H1 nor H2 has abutting group

226.1.1.1, and R2 and R3 accept been absolutely pruned from the (10.1.1.10, 226.1.1.1) SPT. Referring

to Figure 17-10, R1’s s0/0 interface has been pruned from the (S,G) SPT, so R2 and R3 are not

receiving the multicasts beatific by server S1 to 226.1.1.1. The archetype afresh begins with host H2

joining accumulation 226.1.1.1 again.

Figure 17-10 R3 and R2 Accelerate Affix Messages

Without the Affix message, host H2 would accept to delay for as abundant as 3 account afore it would

receive the accumulation traffic. However, with the afterward steps, as listed in Figure 17-10, H2 will

receive the packets in aloof a few seconds:

1. Host H2 sends an IGMP Accompany message.

Fa0/0

10.1.1.0/24

10.1.2.0/24

10.1.5.0/24

10.1.6.0/24

10.1.3.0/24

S0/0 S0/0 .2

S0/1

S0/0

R1 R2

.1 .2

.10

My (10.1.1.10, 226.1.1.1)

entry lists S0/0 and 10.1.3.2

as my RPF interface and

neighbor – accelerate a Affix there!

Graft

My (10.1.1.10, 226.1.1.1)

entry lists S0/0 and 10.1.2.1

as my RPF interface and

neighbor – accelerate a Affix there!

IGMP Join

for 226.1.1.1

2. R3 looks for the RPF interface for its (S, G) accompaniment advice for the accumulation 226.1.1.1 (see

earlier Archetype 17-5), which shows the admission interface as s0/0 and RPF acquaintance as

10.1.3.2 for the group.

3. R3 sends the Affix bulletin out s0/0 to R2.

4. R2 now knows it needs to be accepting letters from 10.1.1.10, beatific to 226.1.1.1. However,

R2’s (S,G) access additionally shows a P flag, acceptation R2 has pruned itself from the SPT. So, R2 finds

its RPF interface and RPF acquaintance IP abode in its (S,G) entry, which references interface

s0/0 and router R1.

5. R2 sends a affix to R1.

At this point, R1 anon puts its s0/0 aback into the approachable interface list, as does R2, and now

H2 receives the multicast packets. Note that R1 additionally sends a Affix Ack bulletin to R2 in response

to the Affix message, and R2 sends a Affix Ack in acknowledgment to R3’s Affix bulletin as well.