DD2490 p4 2010

IP Multicast routing

Multicast routing

Olof Hagsand KTH CSC

1
Literature

•RFC 4601 Section 3 (you may need some definitions from

Section 2). See reading instructions on web.

2
 Deployment

•Multicast routing is in general not deployed in current networks

•However multicast for control is widely used
–Routing and other protocols use multicast locally
•Some sites, eg ”stadsnät” and ISPs have deployed local multicast
delivery
–IPTV distribution may be the killer application
–But this is almost always within a network
•IP Multicast is slowly gaining acceptance

3
IPTV - today
•A multicast driving force today is IPTV - a component of
”triple-play”
You can also distribute video via unicast: http (eg youtube),
p2p (eg bit-torrent), streaming(eg rstp),...
•An operator has a head-end from where it distributes
video content
Terestial, Satelite, hard-disk,....
•The operator distributes the video using IP multicast from
its head-end via its distribution network using IP multicast
routing. Typically based on MPLS.
•From the routers, an L2 network, via xDSL modems,
distributes the video the last step.
•Video is encoded using MPEG-2 (ca 3-4Mb/s) or MPEG-4
(HD 12-16Mb/s)
•Encapsulation is either MPEG directly over UDP (DVB) or
MPEG over RTP over UDP
•One video channel per multicast group
When you change channel, you change group

4
 IP Multicast Overview

•Receiver-based multicast:
–Senders send to any group
–Receivers join groups
•Dynamic group membership
–Hosts leave and join groups dynamically
•Group addresses (class D)
•Uses multicast in hardware if available
•Best-effort delivery semantics (unreliable)
•Notation:
–(S, G) - specific sender S to group G
–(*, G) - all senders to group G
•Prime architect: Steve Deering

5
 IP Multicast Addresses
•IP-multicast addresses, class D addresses (binary prefix: 1110/4)
224.0.0.0 - 239.255.255.255
•28 bit multicast group id
•Selected addresses reserved by IANA for special purposes:

Address Description
224.0.0.0 – 224.0.0.255 Local Network Control Block (dont
forward)
224.0.0.1 All Systems on this subnet
224.0.0.2 All Routers on this subnet
224.0.0.4 DVMRP Routers
224.0.0.9 RIP Routers

233/8 GLOP addressing

232/8 Source-specific multicast

6
 IP Multicast Architecture
3. Host-to router protocol 4. Intra-domain Multicast
Routing

1. Multicast in hosts

2. Link-level Multicast

5. Inter-domain Multicast
Routing

7
 Multicast router

•Listens to all multicast traffic and forwards if necessary.

•Multicast router listens to all multicast addresses.
–Ethernet: 223 link layer multicast addresses
–Listens promiscuously to all LAN multicast traffic
•Communicates with directly connected hosts via IGMP
•Communicates with other multicast routers with multicast routing
protocols

IGMP Multicast
Routing

8
 Multicast in routers
•Packets are replicated to many
output ports
•Forwarding used to be done in slowpath – in CPU
the main CPU
•Modern routers can do forwarding in the
Backplane
linecards or in hardware
•Replication can be made in:
–incoming linecards,
–backplane LC 1 LC 2 LC 3
–outgoing linecards
•Routes computed by main CPU by multicast
routing protocol are installed in the linecard FIB
Network
•Route lookup in two steps:
–RPF,
–multicast forwarding
•In multicast forwarding you look at both the
destination and source IP address

9
 Multicast Routing
•A packet received on a router is forwarded on many interfaces
•The network replicates the packets – not the hosts
•All routing protocols aim at building delivery trees through a
network
•Mostly multicast routing is more concerned where packets come
from instead of where they are going

10
Delivery Trees

•Group Shared Trees: (*, G)

Same delivery tree for all senders
Router state: O(G)
But suboptimal paths and delays
Unidirectional or Bidirectional
Tree's root is in Rendez-vous point (RP)
Also called RP Trees.
•Source Based Trees: (S, G)
Different delivery tree for each sender
Router state: O(S*G),
Optimal paths and delay.
•Data-driven / Push
Trees built when data packets are sent
•Demand-driven / Pull
Build when members join

11
Protocol classification
•Dense-mode protocols
–Intended for domains with high receiver density
–Push, Source trees
–Examples: PIM-DM, DVMRP
•Sparse-mode protocols
–Intended for domains with small receiver density
–Pull, shared trees (not always,..)
–Examples: PIM-SM, CBT
•Link-state protocols
–Source trees
–Examples: MOSPF

12
 Multicast forwarding step 1: RPF
•Reverse Path Forwarding
•Basic forwarding principle in multicast
•Forward datagram only if it arrives on interface used to send unicast
to source
or rendez-vous-point (RP) if (*, G) entry
Send out on all other interfaces: Flooding!
•You can re-use Unicast routing tables for multicast
but used in reverse
•RPF can also be used in unicast to detect source address spoofing
A common attack mechanism

shortest
path
not shortest
path

packet forwarded packet is dropped

13
 Multicast forwarding step 2

•Keep a table with (source, group) entries and an interface

forwarding list
•This is the delivery tree built by the multicast routing protocol
•Data driven (push): Entries are created when data is sent
•Demand-driven (pull): Entries appear when receivers join.

Sender, group interface list

A C *, 231.2.3.4 A, B, C
10.0.0.1, 231.2.3.4 A
B
10.0.2.1, 231.2.3.4 C
*, 229.5.6.7 B, C

14
Exercise1: multicast forwarding
Prefix Outgoing IF
180.70.65.0/24 A
201.4.22.0/24 C
120.14.96.0/24 B
153.18.16.0/24 D
0.0.0.0/0 A
•A router has interfaces A, B, C, D
•The router has a unicast forwarding table and a multicast
forwarding table
•How are the following IP packets forwarded by the
router?
A packet with src = 180.70.65.12 and dst = 228.3.2.2,
arriving on interface B
A packet with src = 201.4.22.4 and dst = 228.3.2.2, arriving
on interface C
A packet with src 120.14.96.42 and dst = 241.16.53.2,
arriving on interface B
A packet with src = 153.18.16.3 and dst = 234.12.32.5.
arriving on interface D Sender, Group Outgoing IF
list
•Solution on web after the lecture *, 228.3.2.2 A, B, C
201.4.22.4, 228.3.2.2 A, D
*, 234.12.32.5 A, C, D

15
 DVMRP
•Distance-Vector Multicast Routing Protocol
–Based on unicast distance vector (eg RIP)
•DVMRP was used in the first multicast Internet
–The MBone.
–Unicast tunnels connected multicast islands.
–Mainly academic use
•DVMRP is data-driven/push and uses source-based
trees
•DVMRP builds Truncated Broadcast Trees
–Reverse Path Broadcasting
–Builds routing tables with source networks
–Uses poison reverse to detect down-link routers
•DVMRP floods trees with data
–Then prunes and grafts tree depending on receiver
dynamics.

16
 Link-State Multicast: MOSPF
•Add multicast to a given link-state routing protocol
•Extend LSAs with group-membership LSA (type 6)
–Only containing members of a group
•Uses the link-state database in OSPF to build delivery
trees
–Every router knows the topology of the complete network
–Least-cost source-based trees using metrics
–One tree for all (S,G) pairs with S as source
•The delivery trees are optimal, since we use Dijkstra
•Expensive to keep all this information
–Cache active (S,G) pairs
–MOSPF is Data-driven/push: computes Dijkstra when
datagram arrives
•If OSPF is used for unicast routing, it is easy(in
principle) to extend it for multicast routing as well

17
 N14

MOSPF example: optimal source-based tree N13

8
3 8
N1 RT1 0 8
RT4 RT5 N12
0
0 N3
source
3 RT2 6
N2 1 6
RT3 RT6
2 7

N4 Ib

Group N15
member Ia
s 2
5 9
RT10 RT7
3
3 1
N11 RT9 0
0
1 0
N9 RT11 N8 N
0
6 0
H1 1 RT12
0
2 RT8
4
N10
N7
18
Protocol Independent
Multicasting - PIM
•PIM relies on any unicast routing tables for its RPF
check
–Does not build its own tables (as DVMRP)
•Split into two protocols for different uses
–PIM-DM – PIM Dense mode
–PIM-SM – PIM Sparse mode

19
PIM – Dense Mode
•PIM - Dense Mode
–Flooding-and-prune strategy – similar to DVMRP
–Creates (S,G) in every router (even if no receivers)
–Ca 3-minute cycle
•Uses unicast routing tables for RPF
–No separate multicast routing protocol
•Builds only source-based trees
•PIM Hello messages – to detect neighbor PIM routers
•PIM Prune messages
–Like DVMRP, but prunes also used instead of poison
reverse in DVMRP
•PIM Assert messages
–To elect one single forwarder on a shared link
•PIM Graft – just like DVMRP

20
 PIM-SM

•Protocol-Independent Multicast - Sparse Mode

•Start with a shared tree
•When receivers join (via IGMP) PIM-Joins travel up to rendez-vous
point to join a shared tree
•When sources start sending traffic, the sources are registered
with the RP
•When new source traffic reaches receiver, a source based-tree
is built
Source-base tree switchover
•How do routers find the Rendez-vous point (RP)?
RP discovery
•There is also a bidirectional variant of PIM-SM

21
 PIM-SM join example

Receiver
1
RP

IGMP join
Sender S
(*, G) Join

Receiver 1 joins the shared

tree

22
 PIM-SM join example

Receiver
1
RP

Sender S

(*, G) Join

IGMP join

Receiver 2 joins the shared

tree
Receiver
2
23
Registering sources – PIM Register
•PIM-SM trees are unidirectional, sender's DR must send to RP to
distribute traffic.
–All routers must know the Group-to-RP mapping
•In bidirectional PIM, a sender can just reach any router in the tree
•Sender's DR sends PIM Register messages towards the RP
–Multicast data is encapsulated within the unicast Register message
•The RP then joins the source tree!
–Sends a (S,G) Join towards the source
•When the native multicast packets arrive at RP:
–Encapsulated data is discarded
•PIM register-stop is sent to sender DR

24
 PIM-SM source registring (1)
•PIM-register sent to RP from source
•Data is encapsulated in unicast Register messages to
RP, then natively in (S,*) tree

1.
Register
New
Sender S

25
 PIM-SM source registring (2)
•RP joins (S,G) tree (hop-by-hop)
•Data is sent natively from S along (S,G) tree to RP and
via Register messages

2. (S, G) Join

2. (S, G) Join
New
Sender S

26
 PIM-SM source registring (3)
•RP sends Register-Stop to make sender stop
encapsulating data.
•Data is now sent natively from S along (S,G) tree to
RP.

New
3. Register-
Sender S
Stop

27
PIM-SM shortest path tree switchover

•For performance reasons, PIM-SM tries to build source-trees (SPT)

after the shared tree join
–Most implementations: try to build source tree directly
•As soon as a receiving router (a DR) gets its first multicast data
packet from a new source S, it sends an (S, G) Join upwards
towards the source.
•To prune the (*,G) tree, it sends a (S,G,rpt) RP-prune at the same
time (to avoid duplicates).
–The RP-prune is a special kind of prune
–The router may still be interested in other (*,G) traffic.
•Finally, the source may be pruned from the shared tree.

28
 PIM-SM switchover(1)

•Receiver DRs sends (S,G) Join towards source joining

the (S,G) tree.
Receiver

Sender S (S, G) Join

Receiver 29
 PIM-SM switchover (2)
•Traffic starts flowing down SPT also
•Receiver DR:s send (S,G,rpt) Prune towards RP to
leave RPT tree and avoid duplicates
Receiver

Sender S (S, G) RP-

Prune

30
Receiver
 PIM-SM Switchover(3)
•Traffic flows down (S,G) tree only
•Note traffic still flows to RP (for new receivers joining
RPT)
Receiver

Sender S

31
Receiver
Source-specific joins

•If a router receives an IGMP v3 source-specific join

•The router can skip the RPT join and go directly to the SPT join!
•The RP and shared tree mechanism is really just a synchronization
mechanism for senders of a group to make contact with receivers
of that group,...

32
PIM Hello protocol

PIM has a Hello protocol

PIM Hellos are sent on 224.0.0.13 to shared media in order to
–Form neighbor adjacencies
–Select DR on a shared media (for handling hosts on a network)
–Options negotiation.

33
Designated Router
•Many multicast routers on one link. IGMP elects one Querier
•But multicast protocol also elects a Designated Router on each
shared segment.
•DR handles multicast routing on behalf of the link, and forwards
data.
•(Note IGMP Querier and DR may be different)

Delivery
tree

DR

34
 PIM Assert

•On a transit link (shared media) one router must be the upstream router
for a tree
•Routers may have inconsistent unicast routing tables causing different
routers to believe different routers are upstream
•This would cause the tree to have several upstream routers ->
duplicates.
•PIM Assert messages are sent to vote for one single forwarder per tree
for the shared media.
•All routers on that shared link use the winner to send joins to.

35
 PIM Rendez-vous points

•The placement of the RP can be important, since all (at least

initial) traffic need to pass the RP.
•Group-to-RP mapping
Same RP for all groups
Different RP's for different groups
•Manual configuration
•Anycast-RP
Use anycast address to identify RP
•But then the RPs need to communicate using
the Multicast Source Discovery Protocol (MSDP)

36
 MSDP – Multicast Source
Discovery Protocol
•Discovering sources is a major problem in PIM, especially in the
inter-domain routing case:
–One AS manages an independent RP for that domain – following the
autonomous systems architecture
–But what if there are senders in other domains?
•MSDP operation:
–The RPs in different domains communicate with each other on a
peer-to-peer basis
–The RP keeps track of all senders in one domain and sends MSDP
Source Active (SA) messages to the RPs in the other domains
–If an RP has receivers in a group G, and discovers a sender S in
another domain, it performs a (S,G) Join towards S.
–SA messages are reliably flooded to all RP peers
• Concern for scalability

37
 MSDP Example(1)

AS2
AS1
RP1 RP2 RP5 AS5
SA(S,G) SA(S,G)
R

SA(S,G) SA(S,G) SA(S,G)

SA(S,G)
S RP3
RP4
AS3
AS4

•Assume a group G and a sender S in AS3 and receiver R in AS5

•S starts sending to G and registers with RP3 in AS3 and R has joined
the shared tree in AS5
•RP3 floods Source Active messages for S
38
 MSDP Example(2)

AS2
AS1
RP1 RP2 RP5 AS5
Join(S,G)
R
Dat
Join(S,G) a

S RP3
RP4
AS3
AS4

•RP5 receives the SA (S,G), and since it has receivers in its domain,
•It joins the source-specific tree (S,G)
•Data flows from S to R via RP5, and then switchover to source-specific
tree
39
Anycast-RP and MSDP

•In anycast-RP the RP is an anycast address which is used by many

routers.
–Every RP serves a part of the network.
•But the RP:s need to synchronize their state to detect all
senders/receivers.
•Anycast-RP:s synchronizes their state with MSDP
–also within a domain.
–This is common practice and used in the lab

40
Exercise 2: Anycast-RP, PIM and MDSP
•Four routers running PIM-SM, Anycast-RP and
MSDP
•RP1 and RP2 are RPs and run MSDP
•A sender S1 sends multicast traffic to group G
IGMP 1
S1
•A receiver R1 receives traffic. 2 DR2

•List the multicast forwarding entries RP1

PIM-SM
1
1 MSDP
Before tree switchover 2
2 RP2
After tree switchover PIM-SM
DR1 1
•Solution on web after the lecture. IGMP
R1
2

Router (Sender, Group) Outgoing i/f list

41
 Source-specific multicast
•Regular IP multicast model (Any-source multicast)
–A receiver R joins a group G (*, G)
–R receives all traffic destined to G
–Senders unknown a priori -> The routing protocol must
detect senders
•Source-specific multicast
–A receiver R joins a group G for sender S only
–R receives traffic only from S destined to G
–The routing protocol need not detect sender
•SSM uses multicast addresses with prefix 232/8
•Requires changes to socket API
•IGMPv3
•But detection of senders not necessary

42
Homework 3
•Similar to Exercise 2.
•Define multicast forwarding state before tree switchover
•H1 sends, H2 and H3 receives. R3 and R4 are RPs running
Anycast-RP and MSDP
•Deadline (for bonus): Next lecture (Mon April 26 10:15)

R1 R2
1 1 2
H1
1 R3 1 R4
2
2
3 1 2

R6 1 R5
4 3
2
R7 3
1
2
3

H2 H3

43