Chapter 16b

1
Chapter 16b

• Multicasting

2
Multicasting Applications

• Multimedia
• television, presentations, etc.
• Teleconferencing
• voice and video
• Database
• replication and updates
• Distributed computing and real-time workgroup
• exchange of results, files, graphics, messages,
  etc.

3
Multicast Example - Broadcast
(or Flooding)
4
Multiple Unicast Example
5
Multicast Strategies
6
True Multicast

• Source knows network location of multicast group
  members
• identifies least cost path to each member
• establishes a partial spanning tree to reach all
  group member networks
• Source node sends single packet along tree path
• Packet is replicated by routers at each branch
  point in path

7
True Multicast Example
8
Multicast Operation
9
Multicast Strategies
10
Multicast Requirements

• Unique multicast addressing scheme
• IPv4 Class D addresses 1110, followed by 28-bit
  group identifier
• IPv6 11111111, 3 flag bits, 4 scope bits, and
  112-bit group identifier
• Nodes must be able to translate between multicast
  addresses and list of networks that have members
• Routers must translate between multicast address
  and subnetwork addressing (e.g. MAC multicast)
• Must have mechanism for hosts to inform routers
  of group membership/exclusion
• Routers must have mechanism for determining and
  applying multicast routing paths (multiple
  outbound paths for the same received packet)

11
Multicast Operation - Duplication Avoidance
How do you avoid this problem?
12
Multicast Operation - Duplication Avoidance

• Multicast routing is based on knowledge of source
  as well as multicast destination
• Each router must calculate spanning tree for
  source node and route on that basis
• Multicast routing algorithms must support this
  capability

13
Internet Multicast Service Model
128.59.16.12 (m/c group member)
128.119.40.186 (source)
multicast group 226.17.30.197
128.34.108.63 (m/c group member)
128.34.108.60 (m/c group member)

• multicast group concept use of indirection
• hosts addresses IP datagram to multicast group
• routers forward multicast datagrams to hosts that
  have joined that multicast group

14
Multicast groups

• class D Internet addresses reserved for
  multicast
• host group semantics
• anyone can join (receive) multicast group
• anyone can send to multicast group
• no network-layer identification to hosts of
  members
• needed infrastructure to deliver mcast-addressed
  datagrams to all hosts that have joined that
  multicast group

15
Multicast Routing Mechanisms

• Group-Shared or Center-Based Tree
• Single optimum (shortest path) shared tree for
  all senders/ receivers
• Typically based on use of a center (rendezvous
  point) of the tree
• Source-Based Tree
• Optimum routing tree for each source (sender) in
  a multicast group
• Typically uses reverse path forwarding

16
Center-Based Routing Trees

• Single delivery tree shared by all
• One router identified as center of tree,
  designated the rendezvous point

1. edge router sends unicast join-msg addressed to
   the rendezvous point (center router)
2. join-msg processed by intermediate routers and
   forwarded towards center
3. join-msg either hits existing tree branch for
   this center, or arrives at center
4. path taken by join-msg becomes new branch of tree
   for this router

17
Center-Based Routing Example
Suppose R6 chosen as center
LEGEND
R1
router with attached group member
R4
3
router with no attached group member
R2
2
1
R5
path order in which join messages generated
R3
1
R7
R6
18
Source-Based Trees with Reverse Path Forwarding

• Rely on routers knowledge of unicast shortest
  path from it to sender
• Each router has simple forwarding behavior

if (mcast datagram received on incoming link on
shortest path back to sender) then flood
datagram onto all outgoing links
else ignore datagram
19
Reverse-Path Forwarding Example
S source
LEGEND
R1
R4
R2
R5
R3
R7
R6
20
Reverse-Path Forwarding Example

• forwarding tree contains subtrees with no
  multicast group members
• no need to forward datagrams down subtree
• prune messages are sent upstream by router with
  no downstream group members

LEGEND
S source
R1
router with attached group member
R4
router with no attached group member
R2
P
P
R5
prune message
links with multicast forwarding
P
R3
R7
R6
21
IGMPv4 Message Format
RFC 2236

• Type
• Membership Query learn group members on network
• Membership Report declare group membership
• Leave Group declare departure from group
• Max Response Time
• in Membership Query only
• max time before sending response in 1/10 second
  units
• Checksum 16-bit ones complement
• Group Address IP multicast address (zero in
  request message)

22
IGMP Operation

• Hosts send Membership Report to join groups
• sent in IP datagram with IP destination address
  equal to Group Address
• therefore, routers and other hosts are informed
  of new member
• Multicast routers periodically broadcast
  Membership Query to maintain current list
• hosts must reply with Report for each group in
  which it wants to remain
• if a host hears another Membership Report for one
  of its groups within a random timer value (lt
  maximum response time) , it cancels its report
  (why??)
• Hosts send Leave when it leaves a group
• Routers then use group-specific Query to
  determine if any other member of this group are
  left

23
IPv6 Group Membership

• IGMP-like functions incorporated into new version
  of ICMP for IPv6 (ICMPv6)
• ICMPv6 includes group membership query and report
  messages, and a new group membership termination
  message

24
Multicast Routing Algorithms

• DVMRP distance vector source-based with RPF/RPM,
  based on RIP
• MOSPF link-state source-based, extension of OSPF
• CBT core-based tree
• PIM-DM protocol independent, dense
• PIM-SM protocol independent, sparse
• MBONE tunneling via backbone

25
Distance-Vector Multicast Routing Protocol (DVMRP)

• The first and, arguably, most widely-deployed
  multicast routing algorithm used in the Internet
• Straightforward implementation of source-based
  trees
• with reverse-path forwarding and pruning
• pruned branches automatically restored after
  specified prune lifetime
• Uses distance-vector algorithm to determine next
  hop for best path back to the source

26
Multicast Extensions to OSPF

• Direct extension to OSPF unicast routing
• MOSPF is designed to operate within a single AS
  to generate source-specific, pre-pruned,
  least-cost trees for each multicast group
• Multicast spanning trees calculated on demand
  using Dijkstras algorithm
• Routers periodically flood group membership
  information to all other routers in its area
• added to the link-state advertisements that are
  used with OSPF

27
MOSPF Routing
28
Protocol Independent Multicast (PIM)

• More general solution to multicast routing
• Key assumption members of any given multicast
  group are few and widely-dispersed
• Independent of underlying unicast routing
  algorithm
• Uses multiple shortest-path unicast routing
  approach
• Two modes of operation (actually, two separate
  algorithms)
• dense mode intra-AS
• sparse mode inter-AS

29
PIM Protocol Independent Multicast

• not dependent on any specific underlying unicast
  routing algorithm (works with all)
• two different multicast distribution scenarios

• Dense
• group members densely packed, in close
  proximity.
• bandwidth more plentiful

• Sparse
• networks with group members small wrt
  interconnected networks
• group members widely dispersed
• bandwidth not plentiful

30
Consequences of Sparse-Dense Dichotomy

• Sparse
• no membership until routers explicitly join
• receiver- driven construction of mcast tree
  (e.g., center-based)
• bandwidth and non-group-router processing
  conservative

• Dense
• group membership by routers assumed until routers
  explicitly prune
• data-driven construction on mcast tree (e.g.,
  RPF)
• bandwidth and non-group-router processing
  profligate

31
PIM- Dense Mode

• flood-and-prune RPF, similar to DVMRP or MOSPF
  but
• underlying unicast protocol provides RPF info for
  incoming datagram
• less complicated (less efficient) downstream
  flood than DVMRP reduces reliance on underlying
  routing algorithm
• has protocol mechanism for router to detect it is
  a leaf-node router

32
Sparse-mode PIM Operation

• One router is designated as the RP (rendezvous
  point) for each multicast group (center-based
  tree, like CBT algorithm)
• Group destination routers send Join messages to
  RP requesting membership for its hosts in the
  RPs group
• uses a unicast shortest path route selection
  (e.g. RIP)
• reverse of this path becomes part of that RPs
  shared distribution tree
• Source nodes send messages intended for a group
  to the RP for that group
• uses a unicast shortest path route from source to
  RP
• RP routes packets, using the group-shared tree,
  back toward routers that have Joined that group

33
PIM - Sparse Mode

• center-based approach
• Group router(s) send join msg to the rendezvous
  point (RP)
• intermediate routers update state and forward the
  join
• after joining via RP, router can switch from the
  group-shared tree to a source-specific tree
• increased performance less concentration,
  shorter paths

R1
R4
join
R2
join
R5
join
R3
R7
R6
rendezvous point
34
PIM - Sparse Mode

• Multicast sources(s)
• Register with RP
• RP can extend the multicast tree upstream to a
  source
• Send data to RP, via unicast path, which then
  distributes data down RP-rooted tree
• RP can send stop msg to a source if all receivers
  leave a group
• no one is listening!

R1
R4
R2
R5
RP
R3
R7
R6
all data multicast from Rendezvous Point
Source
35
PIM Routing Example
?
?
?
?