Ahmed Helmy Assistant Professor of Computer Engineering ...afn48922/downs/wireless/Multicast_Tutorial.pdf · – DVMRP is a distance-vector routing protocol, – MOSPF is an extension

�� Ahmed Helmy

Assistant Professor of Computer Engineering

University of Southern [email protected]

http://ceng.usc.edu/~helmy

Ahmed Helmy - USC 2

��- Motivation and Background

- Multicast vs. unicast

- Multicast Applications

- Delivery of Multicast- Local delivery and multicast addressing

- WAN delivery and its model

- Group Membership Protocol (IGMP)

- Multicast Algorithms and Concepts- Flooding, Spanning Tree, Reverse Path

Broadcasting (RPB), Truncated RPB, ReversePath Multicasting, Center-Based Trees

Ahmed Helmy - USC 3

- Multicast Routing Protocols- Dense vs. Sparse Multicast

- DVMRP

- MOSPF

- PIM (PIM-DM, PIM-SM)

- Multicast and the Internet- The MBONE

- Recent deployment

- Multicast and IP-Mobility

��

Ahmed Helmy - USC 4

Session 18:30-10:00

• Multicast vs. Unicast

• Multicast Applications

• Multicast Delivery and Host Model

Ahmed Helmy - USC 5

��

• Multicast provides multipoint-to-multipointcommunication

• Today majority of Internet applications rely onpoint-to-point transmission (e.g., TCP).

• IP-Multicast conserves bandwidth by replicatingpackets in the network only when necessary

Ahmed Helmy - USC 6

��S

R1

R2

R3R4

S

R1

R2

R3R4

Multiple unicasts Multicast

Ahmed Helmy - USC 7

• �� – Stock ticker application transmits packets to 100

stations within an organization's network.

• Unicast transmits100 packets where many packetsmay traverse the same links.

• Multicast transmits one packet that is replicateddistributed by the multicast delivery tree.

Ahmed Helmy - USC 8

�� • One-to-Many

– Scheduled audio/video distribution: lectures,presentations

– Push media: news headlines, weather updates– Caching: web site content & other file-based updates

sent to distributed replication/caching sites– Announcements: network time, configuration updates– Monitoring: stock prices, sensor equipment

Ahmed Helmy - USC 9

��

• Many-to-One– Resource discovery

– Data collection and sensing

– Auctions

– Polling

Ahmed Helmy - USC 10

– Many-to-Many• Multimedia Teleconferencing (audio, video, shared

whiteboard, text editor)

• Collaboration

• Multi-Player Games

• Concurrent Processing

• Chat Groups

• Distributed Interactive Simulation

��


- Currently: The Multicast Backbone(MBONE) carries audio and videomulticasts of IETF meetings, NASA spaceshuttle missions,.. etc.


� �� ! �� " �� #- The Session Directory (SD) tool lists active

multicast sessions on MBONE and allowsto join a conference using MBONE tools:- vat (visual audio tool), rat (robust audio tool)

- vic (video tool)

- wb (shared white board)

- nte (network text editor), .. etc.






• Resource Discovery– Multicast may be used (instead of broadcast) to

transmit to group members on the same LAN.

– Multicast may be used for resource discoverywithin a specific scope using the TTL field inthe IP header.

� ��


��$� ��% �� &TTL Expanding-Ring Search

to reach or find a nearby subset of a group

s

1

2

3


��$� ��% �� &Administrative TTL Boundaries

to keep multicast traffic within anadministrative domain, e.g., for privacyreasons

an administrative domain

TTL threshold set oninterfaces to these links,greater than the diameterof the admin. domain

the rest of the Internet


��$� ��% �� &Administratively-Scoped Addresses

an administrative domain

address boundary set oninterfaces to these links

the rest of the Internet

– RFC 1112

– uses address range 239.0.0.0 — 239.255.255.255


• Over the same (LAN):– The source addresses the IP packet to the

multicast group

– The network interface card maps the Class Daddress to the corresponding IEEE-802multicast address

– Receivers notify their IP layer to receivedatagrams addressed to the group.

– Key issue is ‘addressing’ & filtration

'�� (��)��* +��(��


• Over different subnets:– Routers implement a multicast routing protocol

that constructs the multicast delivery trees andsupports multicast data packet forwarding.

– Routers implement a group membershipprotocol to learn about the existence of groupmembers on directly attached subnets.

– Hosts implement the group membershipprotocol that provides the ‘IP-multicast hostmodel’


��

• Types of IP addresses:– Unicast: used to transmit packets to one destination.

– Broadcast: used to send datagrams to entire subnet.

– Multicast: used to deliver datagrams to a set ofhosts (members of a multicast group) in variousscattered subnets.


• IP-Multicast is a “best-effort” service.– Reliable/ordered delivery are not guaranteed.

– Reliability may be provided by upper-layerprotocols (e.g., reliable multicast protocols).

• IP-Multicast packets include a "group address"(Class D) in the Destination field of the IPheader.


��

• An IP multicast group is identified by aClass D address.

• Multicast group addresses range from(224.0.0.0) to (239.255.255.255).


• The Internet Assigned Numbers Authority(IANA) registers IP multicast groups.

• The block of multicast addresses rangingfrom (224.0.0.1) to (224.0.0.255) isreserved for local LAN multicast:– used by routing protocols and other low-level

topology discovery or maintenance protocols

– E.g., "all-hosts" group (224.0.0.1), "all-routers”group (224.0.0.2), "all DVMRP routers", etc.

• The range (239.0.0.0) to (239.255.255.255)are used for site-local "administrativelyscoped" applications.


• All multicast addresses in IANA's reservedblock begin with 01-00-5E (hex)

• Mapping between a Class D and anEthernet multicast address is obtained by:– placing the low-order 23 bits of the Class D

address into the low-order 23 bits of IANA'sreserved address block.

��%��(� ��!��


– How multicast group address 224.10.8.5 (E0-0A-08-05) ismapped into an Ethernet (IEEE-802) multicast address.

• The mapping may place up to 32 different IP groupsinto the same Ethernet address because the upperfive bits of the IP multicast group ID are ignored.


• Hosts can join or leave a group at any time

• A host may be a member of multiple groups

• Senders need not be members of the group

• Participants do not know about each other

• The two components of IP-multicast:– the group membership protocol

– the multicast routing protocol

'!��



• Routers need to learn about the presence ofgroup members on directly attached subnets

• When a host joins a group:– it transmits a group membership message for

the group(s) that it wishes to receive

– sets its IP process and network interface card toreceive packets sent to those groups.

,� ��"��!��


• This receiver-initiated join process scaleswell:– as the group size increases, it becomes more

likely for a new member to locate a nearbybranch of the multicast distribution tree.


- �� – Run on routers and establish the multicast

distribution tree to forward packets from sender(s)to group members.

• Based on unicast routing concepts:– DVMRP is a distance-vector routing protocol,

– MOSPF is an extension to the OSPF link-stateunicast routing protocol.

• Center-based trees (e.g., CBT & PIM-SM)introduce the notion of the tree ‘core’.


Session 210:15-12:00

• IGMP

• Multicast Algorithms


��,� �� ,��

• IGMP runs between hosts and theirimmediately neighboring multicast routers.

• The protocol allows a host to inform its‘first-hop’ router that it wishes to receivepackets destined to a specific group.


• Routers periodically query the LAN todetermine if group members are still active.

• One router per LAN is elected as "querier"to query for group members.

• Through IGMP a router determines whichmulticast traffic needs to be forwarded toeach of its "leaf" subnets.

� ��.�� ,��


�,��/�� 0

• RFC-1112

• To determine local group membership:– Multicast routers periodically transmit ‘Host

Membership Query’ messages

– Queries are ddressed to the all-hosts group(224.0.0.1) with TTL = 1 (i.e., not forwardedby any other multicast router).


• Upon receiving a Query, a host responds with a‘Host Membership Report’ for each group that itwishes to Join

• Observation: The router only needs to know of atleast one group member on the leaf subnet

��1 ��,� ��


�� $�� !��

• To avoid Report implosion:– Each host starts a random delay timer for its Reports.

– If during the delay period another Report is heard forthe same group, the host resets its timer

– Otherwise, the host transmits a Report causing othergroup members to reset their timers

• Thus, Reports are spread out over time and Reporttraffic is minimized


– IGMP-Query Message


• The querier periodically transmit Queries toupdate local membership

• If no Report is received for a group after anumber of Queries, the router assumes thatmembers are no longer present on that LAN– the group is removed from the membership list

of that interface/subnet

2�� "��!��


��1 ��*

• When a host first joins a group, itimmediately transmits a Report for thegroup rather than waiting for a routerQuery.


• IGMP V2 was part of IP-mcast (V3.3-3.8)– spec <draft-ietf-idmr-igmp-v2-01.txt>

• IGMP V2 enhances IGMP V1– IGMP V2 elects one querier for each LAN, the

router with the lowest IP address.

– In V1, the querier election was done by themulticast routing protocol (different multicastrouting protocol used different methods).

��,��/3�


• IGMP V2 defines a new Query message,the ‘Group-Specific Query’, to Query aspecific group rather than all groups


• To reduce ‘leave latency’ V2 defines a ‘LeaveGroup’ message– When a host leaves a group, it sends a ‘Leave Group’

to the all-routers group (224.0.0.2) with the group fieldset to the group to be left.

– Upon receiving a Leave from a LAN, the queriersends Group-Specific Query on that LAN.

– If there are no Reports in response to the Group-Specific Query, the group is removed from themembership list of that subnet.

��)��*


• Observation: With IGMP V1 and V2, if ahost wants to receive any sources from agroup, the traffic from all sources for thegroup has to be forwarded onto the subnet.


• Spec: <draft-ietf-idmr-igmp-v3-03.txt>

• IGMP V3 supports Group-Source Reports:– A host can elect to receive traffic from specific

sources of a multicast group.

– An inclusion Group-Source Report specifies thesources a host wants to receive.

– An exclusion Group-Source Report identifiesthe sources a host does not want to receive.

��


• IGMP v3 enhances support for LeaveGroup messages to support ‘Group-Source’Leave messages:– A host can leave an entire group or specific

(source, group) pair(s).


��4 �� !��

• A multicast routing protocol is responsiblefor the establishment of the multicastdistribution tree and for performing packetforwarding.


• Several algorithms may be employed bymulticast routing protocols:• Flooding

• Spanning Trees

• Reverse Path Broadcasting (RPB)

• Truncated Reverse Path Broadcasting (TRPB)

• Reverse Path Multicasting (RPM)

• Core-Based Trees


• These algorithms are implemented in themost prevalent multicast routing protocolsin the Internet today.• Distance Vector Multicast Routing Protocol

(DVMRP)

• Multicast OSPF (MOSPF)

• Protocol-Independent Multicast (PIM) [PIM-DM and PIM-SM]


4� ��• The simplest technique for multicast delivery.

• When a router receives a multicast packet itdetermines whether or not this is the first timeit has seen this packet.– On first reception, a packet is forwarded on all

interfaces except the one on which it arrived.

– If the router has seen the packet before, it isdiscarded.


• A router does not maintain a routing table, butneeds to keep track of recently seen packets.

• Flooding does not scale for Internet-wideapplication:- Generates a large number of duplicate packets and

uses all available paths across the internetwork.

- Routers maintain a distinct table entry for eachrecently seen packet (consumes memory).


5$��'��• More effective than flooding

• Defines a tree structure where one active pathconnects any two routers on the Internet.

• Spanning Tree rooted at R


• A router forwards each multicast packet tointerfaces that are part of the spanning treeexcept the receiving interface.

• A spanning tree avoids looping of multicastpackets and reaches all routers in thenetwork.


• A spanning tree algorithm is easy toimplement

• However, a spanning tree solution:– may centralize traffic on small number of links

– may not provide the most efficient pathbetween the source and the group members.


��)��!6� ��6�

• More efficient than building a singlespanning tree for the entire Internet.

• Establishes source-rooted distribution treesfor every source subnet.

• A different spanning tree is constructed foreach active (source, group) pair.


��6�� !�• For each (source, group) pair

– if a packet arrives on a link that the routerconsiders to be the shortest path back to the sourceof the packet

• then the router forwards the packet on all interfacesexcept the incoming interface.

– Otherwise, the packet is discarded.


• The interface over which a router acceptsmulticast packets from a particular source iscalled the "parent" link.

• The outbound links over which a routerforwards the multicast packets are called the"child" links.


• Reverse Path Broadcasting (RPB)Forwarding


• Enhancement to reduce packet duplication:– A router determines if a neighboring router

considers it to be on the shortest path back tothe source.

– If Yes, the packet is forwarded to the neighbor.

– Otherwise, the packet is not forwarded on thatpotential child link.


• To derive the parent-child information:– link-state routing protocol already has it (since

each router maintains a topological database forthe entire routing domain).

– distance-vector routing protocol uses ‘poisonreverse’:

• a neighbor can either advertise its previous hop forthe source subnet as part of its routing updatemessages or "poison reverse" the route.


• �� +��)��!6� ��


6��+��

• Reasonably efficient and easy to implement.

• Does not require keeping track of previouspackets, as flooding does.

• Multicast packets follow the "shortest" pathfrom the source to the group members.

• Avoids concentration over single spanning tree


��

• Does not take into account group membershipwhen building the distribution tree.

• As a result, packets may be unnecessarilyforwarded to subnets with no group members.


'��)��!6� ��'��6�

• Using IGMP, routers discover groupmembers and avoid forwarding packets ontoleaf subnets with no members.

• The spanning delivery tree is "truncated" ifa leaf subnet has no group members.


• Truncated Reverse Path Broadcasting(TRPB)


• TRPB eliminates unnecessary traffic on leafsubnets

• But it does not consider group membershipwhen building the branches of thedistribution tree.


��)��!��

• RPM enhances TRPB.

• RPM creates a delivery tree that spans only:-Subnets with group members-Routers and subnets along the shortest pathto group members

• In RPM, non-member branches are pruned

• Packets are forwarded only along branchesleading to group members.


��5.��

• The first multicast packet is forwarded(using TRPB) to all routers in the network.

• Routers at edges of the network with nodownstream routers are called ‘leaf’routers.

• A leaf router with no downstream memberssends a "prune" message on its parent linkto stop packet flow down that branch.


• Prune messages are sent hop-by-hop backtoward the source.

• A router receiving a prune message storesthe prune state in memory.

• A router with no local members thatreceives prunes on all child interfaces sendsa prune one hop back toward the source.

• This succession of prune messages creates amulticast forwarding tree that contains onlybranches that lead to group members.


• Reverse Path Multicasting (RPM)


• To adapt to membership/network dynamics,the prune state is timed out periodically, andpackets are broadcast throughout the network.

• This may result in a burst of prune messages.


��

• Despite improvements over RPM, there arescaling issues and limitations:– Multicast packets are periodically forwarded to

every router in the network.

– Routers maintain prune state off-tree for all(source,group) pairs.

• These limitations are amplified with increasein number of sources and groups.


%��7% ��6��'��%6'�

• Earlier algorithms build source-based trees

• CBT builds a single delivery tree (rooted at thecore) that is shared by all group members.

• Multicast traffic for each group is sent andreceived over the shared tree, regardless of thesource.


• A core-based tree involves one or morecores in the CBT domain.

• Each leaf-router of a group sends a hop-by-hop "join" message toward the "core tree"of that group.

• Routers need to know the group core tosend the join request.

%6'.��


• Intermediate routers process the join request:– The interface on which the join was received is

added to the delivery tree.

• Intermediate routers forward join requeststoward the core until the join reaches a core ora router on the distribution tree.

• Senders unicast their packets toward the core.

• When the unicast packet reaches a member ofthe delivery tree, the packet is multicast to alloutgoing interfaces that are part of the tree.


6��+��

• Advantages over RPM, in terms of scalability:– A router maintains state information for each

group, not for each (source, group) pair.

– Multicast packets only flow down branchesleading to members (not periodically broadcast).

– Only join state is kept on-tree


�� • CBT may result in traffic concentration near the

core since traffic from all sources traverses thesame set of links as it approaches the core.

• A single shared delivery tree may create sub-optimal routes resulting in increased delay.

• Core management issues– dynamic core selection

– core placement strategies


Session 31:00-2:15

• Multicast Routing Protocols– Dense vs. Sparse Multicast

– DVMRP

– MOSPF


��

• In general, there are two classes of multicastrouting protocols:– Dense-mode protocols (broadcast-and-prune)

• DVMRP, PIM-DM, (MOSPF!)

– Sparse-mode protocols (explicit-join)• PIM-SM, CBT, BGMP


(��)��$��

R1

R2

R3R4

S

Dense-Mode Multicast


(��)��$��

S

R1

R2

R3R4

RootR1

R2

R3R4

S

Dense-Mode Multicast Sparse-Mode Multicast


(��/�� (/��

• DVMRP constructs source-rooted treesusing variants of RPM.

• Many MBONE routers run DVMRP

• DVMRP was first defined in RFC-1075.– The original spec was derived from the Routing

Information Protocol (RIP) and used TRPB.

– Mrouted version 3.8 uses RPM.


• Physical/Tunnel Interfaces and Scoping:– The ports of a DVMRP router may be either

• a physical interface to a directly attached subnet or

• a tunnel interface to another multicast island.

• Interfaces are configured with a TTL thresholdthat limits the scope of a multicast transmission.

• Each tunnel interface is configured with the IPaddress of the remote router's interface.


• A router forwards a packet only if:– TTLPacket > TTLthreshold of interface

TTL Scope Values


(/��6��.��

• DVMRP implements RPM.

• The first packet for any (source, group) pairis broadcast to the entire network, providingthe packet's TTL permits.

• Leaf routers with no local members sendprune messages back toward the source.


• Prune messages cause the removal ofbranches that do not lead to group members

• The result is source-specific shortest pathtree with all leaves having group members.

• After a period of time, the pruned branchesgrow back and the packets are broadcastthroughout the network.

• A “graft” mechanism helps to quickly re-establish previously pruned branches.


• A new member joining the group causesthe first-hop router to send a graft messageto the group's previous-hop router.

• When an upstream router receives a graftmessage, it removes the prune state.

• Graft messages may cascade back towardthe source allowing previously prunedbranches to be restored.


��(/��$�� g g

s

g


��6� ��'��6� ��

g g

s

g


�� "��"��!��g g

s

prune (s,g)

prune (s,g)

g


graft (s,g)

graft (s,g)

,��+��"��g g

s

g

g

report (g)


(/��(��"�� '��g g

s

g

g


�) �� • To avoid duplicates, one router per LAN is

elected the Dominant Router


• The router with lowest metric to the sourcesubnet (with the lowest IP address as tiebreaker) becomes the Dominant router

• A dominant router is ‘the’ forwarder for theLAN for traffic from the source subnet


• Entries in a typical DVMRP forwarding table:– Source Subnet

– Multicast Group

– InPort - The parent port for the (S, G) pair. A "Pr"indicates that a prune was sent to upstream.

– OutPorts - The child ports over which packets forthe (S, G) pair are forwarded. A ‘p’ indicates prunemessage received on that interface.

(/��4 ��'�"��


(/��4 ��'�"��


�� .$�4��.$�4�

• OSPF V2 is defined in RFC-1583.

• OSPF is a unicast routing protocol thatdistributes topology information andcalculates routes for a single domain.

• MOSPF is defined in RFC-1584.

• MOSPF routers maintain a current image ofthe network topology through the unicastOSPF link-state routing protocol.


• MOSPF does not support tunnels

• Basic MOSPF runs in a single OSPF domain

• MOSPF uses IGMP to discover members ondirectly attached subnets.

• The Designated Router (DR) is responsible forsending membership information to all routers inthe OSPF domain.

• The DR floods Group-Membership Link StateAdvertisements (LSAs) throughout the OSPFdomain.


• The shortest path tree for (S, G) pair is built "ondemand" when a router receives the first packetfor (S,G).

• When the initial packet arrives, the sourcesubnet is located in MOSPF link state database.– MOSPF LS-DB = OSPF LS-DB + Group-Membership LSAs

6��!�$! ��!'��


• Source-rooted shortest-path tree isconstructed using Dijkstra's algorithm.

• To forward packets to downstreammembers, each router determines itsposition in the shortest path tree

• After the tree is built, Group-MembershipLSAs are used to prune those branches thatdo not lead to group members.


• All routers within an OSPF domaincalculate the same shortest path trees.

• MOSPF LS-DB allow a router to performRPM computation "in memory".

• No need for broadcast and prune.


• Forwarding cache entry contains the(source, group) pair, the upstream node, andthe downstream interfaces.

• MOSPF Forwarding Cache

4 ��%��!�


• The forwarding cache contains:– Destination: The group address

– Source: The packet’s source subnet.

– Upstream: The interface from which (S,G)packets are received.

– Downstream: The interfaces to which (S,G)packets are forwarded

– TTL: min. number of hops a packet needs toreach the group members. [This allows therouter to discard packets with no chance ofreaching the members.]


• The forwarding cache is not aged. Theforwarding cache will change when:• The topology of the OSPF network changes,

forcing all of the datagram shortest-path trees tobe recalculated.

• There is a change in the Group-MembershipLSAs indicating that the distribution ofindividual group members has changed.


��

• Limited to OSPF domains

• Flooding membership information does notscale well for Internet-wide multicsat


Session 42:30-3:45

• Protocol Independent Multicast (PIM)– PIM Dense Mode

– PIM Sparse Mode

– RP and Bootstrap Mechanisms


��

• Design Rationale:– Broadcast and prune keeps state off-tree and is

suitable when members are densely distributed

– Explicit join/center-based approach keeps stateon-tree and is suitable when members aresparsely distributed

– PIM attempts to combine the best of bothworlds


(��%! ��

• Shared trees or shortest path trees?– Both: use shared trees to ‘Rendezvous’ then

switch to shortest path to deliver

• DV or LS for routing?– Use routing tables regardless of which protocol

created them (hence the name ‘ProtocolIndependent’)


��.��

• PIM provides both dense-mode (DM) andsparse-mode (SM) protocols

• PIM-DM: similar to DVMRP but does notbuild its own routing table

• PIM-SM: similar to CBT but providesswitching to SPT and bootstrap mechanismfor electing the tree center dynamically


��(�� • Packets initially flow on broadcast tree

• Forwarded away from source using the RPFalgorithm– A router forwards a multicast datagram if received on

the interface used to send unicast datagrams to thesource

• Then, Prunes are sent up the tree to removebranches with no members


��(�� Source

Receiver 2

Receiver 1

D F

I

B

C

A

E

G

H



Prune

Receiver 2Receiver 1

D F

I

B

C

A

E

G

H



Asserts


D F

I

B

C

A

E

G

H




D F

I

B

C

A

E

G

H



Prune


Join Override

Prune

D F

I

B

C

A

E

G

H




D F

I

B

C

A

E

G

H



Graft

Receiver 2

Receiver 3

Receiver 1

D F

I

B

C

A

E

G

H



Receiver 2

Receiver 3

Receiver 1

D F

I

B

C

A

E

G

H


��$��

• A Rendezvous Point (RP) is chosen as tree centerper group to enable members and senders to“meet”

• Members send their explicit joins toward the RP

• Senders send their packets to the RP

• Packets flow only where there is join state

• (*,G) [any-source,group] state is kept in routersbetween receivers and the RP


��$�� • When should we use shared-trees versus source-

trees?– Source-trees tradeoff low-delay from source with

more router state

– Shared-trees tradeoff higher-delay from source withless router state

• Switch to the source-tree if the data rate is abovea certain threshold


��$�� Source

B

E

A D

C

RP


Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP

Source


(*, G)Join

EC Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP

Source


EC Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

Register

EC Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

(S, G)Join

(S, G)Join

EC Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

Register-Stop

EC Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

(S, G) JoinEC Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

(S, G) RP Bit Prune

EC

(S, G) Prune

Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

EC(*, G) Join

Link

(*,G) Data

(S,G) Data

Control


��$��

BA DRP


Source

EC Link

(*,G) Data

(S,G) Data

Control


Host

A B C

1. IGMP Host-

Membership Query

2. IGMP Host-Membership Report for G

3. Create (*,G) entry:Multicast address=G

RP-address=C,WC=1,RP=1outgoing interface list={1}

incoming interface=2

4. Send Join/Prune message to B:

Multicast address=GJoin={C,WC,RP}

Prune=Null

5. Create (*,G) entry:Multicast address=G

RP-address=C,WC=1,RP=1outgoing interface list={1}


6. Send Join/Prune message to C:

Multicast address=GJoin={C,WC,RP}

Prune=Null

D 7. Create (*,G) entry:Multicast address=G

RP-address=C,WC=1,RP=1outgoing interface list={1}incoming interface=Null

1 2 12

3 1...Receiver

LAN PIM DR/IGMPQuerier for LAN

Rendezvous Point (RP)for group G

��)��1 ��!�$!��'��


Host

A

C X

D Host

1. Data packets for G2. Create (S,G) entry


3. Encapsulate Datapackets in Register

messages and unicast to RP(C)

4. Initiate (S,G) packet counter

5. If (*,G) state exists thendecapsulate Registers and

forward packets to oiflist (*,G)

6. If Register data rate > Thresholdthen create (S,G) entry:

outgoing list=oiflist (*,G)-{2}incoming interface=2

RP=0,SPT=0

7. Send Join/Prunemessage to X:

Multicast address=GJoin={S}, Prune=Null 8. Create (S,G) entry:

outgoing list={1}incoming interface=2

9. Send Join/Prunemessage to D:

Multicast address=GJoin={S},Prune=Null

10. Update (S,G) entry:add 2 to outgoing

interface list

11. When receive (S,G) nativepackets set SPT bit for (S,G) entry,

& trigger Register-Stop message to D

12. When receiverRegister-Stop stop

encapsulating packets

Receiver

Source

LAN(B)

DR for LAN(B)

1

2

12

1

2

Rendezvous Point(RP for group G 1

2

��$�� !�,� ��


Host

A B C

1. Receive S’s packets on shared RP treeInitiate packet count

If data rate > Threshold then:Create (S,G) entry:

outgoing interface list={1}incoming interface=2RP=0,WC=0,SPT=0

2. Send Join/Prune message to B:

Multicast address=GJoin={S}

Prune=Null

3. Create (S,G) entry:outgoing interface list={1}

incoming interface=2RP=0,WC=0,SPT=0

6. After receiving packets from D:Set (S,G)’s SPT-bit=1 and,

send Join/Prune message to C:Multicast address=G

Join=NullPrune={S,RP-bit}

D

7. Create (S,G) entry:oif list=oif(*,G)-{1}

RP-bit=1

1 2 12

3 1Receiver

LAN PIM DR/IGMPQuerier for LAN

Rendezvous Point (RP)for group G

4. Send Join/Prune message to D:

Multicast address=GJoin={S}

Prune=Null

5. Add interface 2to the outgoing interface

list of (S,G) entry

2

12

HostSource (S)

First Hop Router for S

$��!�� !�$! ��!'��


'!��6 �� "��• Which router to use as RP for a group?

– A set of well-connected routers are configured asCandidate-RPs for group(s) per domain

– A manageable number of RPs is chosen

– RPs advertise candidacy for group-prefix (notper group), for scalability

– Periodic advertisement of candidacy to capturedynamics and unreachability

• Who maintains/updates/distributes this info?


��6 ��(�� • Host model:

– hosts need only “logical” multicast group addressto send or receive

• RP address is network (not logical) address

• Routers should map group address to RP addressand adapt to unreachability/change of RP


��6 ��(�� • No “on-demand” retrieval of RP info to avoid

start-up phase• can’t join or send until DR gets RP address

• “bursty source” problem:• packets are lost until DR identifies active RP

• global distribution of explicit group to RPmapping and reachability not scalable

• Use a-priori status distribution• like unicast routing, periodic liveness tracking

• distribute RP-list throughout the domain


%! ��&'!�6 ��!��

• PIMv2 has a Bootstrap router election procedure– The Bootstrap router receives Candidate-RP messages

from potential RPs

– Bootstrap router sends Bootstrap messages whichcontain a list of reachable Candidate-RPs

– All PIM routers receive these Bootstrap messages

– DRs obtain group-to-RP mapping (when hosts join orsend to the group) through a hash algorithm


��6 ��!��

• RP location need not be optimized, butconsistent RP mapping and adaptation tofailures is criticial– all routers (within PIM domain) must associate

a single active RP with a multicast group

• Routers use ‘algorithmic mapping’ ofGroup address to RP from manageably-small set of RPs known throughout domain


��6 ��!��• Each candidate RP indicates liveness to the

Bootstrap Router in the PIM domain

• Bootstrap Router distributes set of reachablecandidate RPs to all PIM routers in domain.

• Each PIM router uses the same hashfunction and set of RPs to map a particularmulticast group address to that group’s RP.


(*��6 ��

• Simple bridge-like spanning-tree election algorithm

• A set of well-connected routers are configured asCandidate Bootstrap Routers (C-BSRs) per domain

• C-BSRs originate PIM hop-by-hop Bootstrap messageswith IP address and preference value.

• Bootstrap messages are exchanged by all PIM routerswithin domain (flooded with RPF check)

• Most preferred (or highest numbered) reachable C-BSRis elected


� ��!��!+�� ,� ��

• Hash function– input: group address G and address of each candidate

RP in RP set (with optional Mask)

– output: Value computed per candidate RP in RP set

– RP with highest value is the RP for G

• Desirable characteristics– minimize remapping when RP reachability changes —

remap only those that lost RP

– load spreading of groups across RPs


�� 2��!�"��*• When Candidate RP fails/unreachable

– Bootstrap Router times it out

– Bootstrap message distributed with updated RPset

– Routers hash affected groups to different RP


��+��

• RFC 2362/2117

• http://catarina.usc.edu/pim


Session 54:00-5:00

• Multicast and the Internet– The MBONE

– Recent Deployment

– BGMP


��!��

• Initially there was the MBONE

• Short-term inter-domain solution based onPIM-SM, MBGP and MSDP

• Longer-term architecture BGMP


'!��8��6��" ��6.��

• The MBONE is an interconnect of subnetsand routers that support IP-multicast.

• The goal of the MBONE was:– initially: to construct an IP multicast test-bed

– as it became popular: gradual deployment ofmulticast applications without waiting for theubiquitous Internet multicast deployment


• The MBONE is rapidly growing– 40 subnets in 4 countries in ‘92

– > 2800 subnets in over 25 countries in April ‘96

• The MBONE is a virtual network layered ontop of a subset of the Internet.

• It is composed of islands of multicast-capablerouters connected to other islands by virtualpoint-to-point links called “tunnels.”


- Tunnels allow multicast traffic to passthrough the non-multicast-capable parts ofthe Internet.

- Multicast packets are encapsulated as IP-in-IP, so they look like normal unicast packetsto intermediate routers.

- Encapsulation is added on entry to a tunneland stripped off on exit from a tunnel.

'��


�� !� ��!��


• The MBONE and the Internet have differenttopologies, so:– multicast routers execute a separate routing

protocol to forward multicast packets.

- Much of the MBONE routers run DVMRP

- Portions of the MBONE run:- MOSPF

- Protocol-Independent Multicast (PIM)



�6.�� • “Mbone currently using DVMRP, which was

never intended for, and is ill-suited to, this task– known problems of DV with large networks

– broadcast & prune approach ‘undesirable’ forinterdomain routing”, S. Deering.

• Suggested solution:– Use sparse-mode concepts

– Use 2-level hierarchy (as in unicast)


��(�� *��

• Use PIM-SM as intra-domain multicastrouting protocol

• Use MBGP (Multicast BGP) to distributeinter-domain multicast routes

• Use MSDP (Multicast Source DiscoveryProtocol) between RPs in different domains


�6,�

• BGP (RFC 1771) used for unicast routing to:– aggregate and abstract routes for scalability

– provide inter-domain routing policies

• BGP4+ (RFC 2283) can carry multicast routes– multicast routers need only know

• - internal topology and - paths to reach other domains

– provides topology info for multicast routes thatmay be different than unicast routes


�� "��% ��$��

( ��$�� ( ��6��$��

�� 6$ �

$ ��!��!�� )��9 �� !��!��

��*+ ��)�� 6� �� " �� )��)��


�$(�

• To tie PIM-SM trees in different domains– every RP has MSDP peers (RPs in other domains)

– when a source registers to the RP it conveys this infoto its MSDP peers through TCP and SA messages

– this info is RPF-flooded to other domains

– an RP with members in its domain joins towards src

RP1

Source S

Last hop routersends (S,G) Register to RP1

RP1 CreatesState

RP2

Receiver R

(S,G)JoinstowardsRP2

AS 2 AS 1

��$�

��

RP1

Source S

RP2

Receiver R

MSDP PeeringMSDP Peering o Between RPs o Over TCP

$��$��

RP1

Source S


RP1 CreatesState

RP2

Receiver R

RP1 Sends (S,G) SA message


MSDP Peering

1 ��!�$ ��'��

RP1

Source S


RP1 CreatesState

RP2

Receiver R

RP2 Joins (S,G) Source Tree


(S,G) Joins

4 ��

RP1

Source S

RP2

Receiver R


��

• Short-term solution that doesn’t scale well!


��(�)�� ( ��

• BGMP (Border Gateway Multicast Protocol):– PIM-SM-like inter-domain multicast routing protocol

– builds bi-directional shared trees of domains

– each tree has a ‘root domain’ (like an RP)

• MASC (Multicast Address Set Claim):– mechanism to associate addresses with root domains

• MBGP:– extends BGP to convey ‘address-range to root’ mapping

to border routers


6,��• Bi-directional shared trees rooted at

domains

• Border routers send joins and data towardroot domain for mcast address in packet

• Mapping of multicast address to rootdomain obtained from BGP4+ MRIB

• Source specific branches only where“needed”


ISP 1

Sender/Rcvr

Group Initiator

BGMP tree

AS1


.�!��!��( ��!��

• Multicast address allocation

• Multicast state aggregation

• Security and authentication (always an issue!!)

• …

• For more references:

– Sigcomm ‘99 [Kumar et al.]– http://catarina.usc.edu/multicast


Session 68:30-10:00

• Mobile IP

• Multicast and IP-Mobility


� "��• Each mobile node has a home network,

home address and home agent

Home Agent (HA)

Home Network

Mobile Node

Correspondent Node


Home Agent

Home Network

Correspondent Node

Foreign Agent (FA)

Foreign Network

Mobile Node

• When mobile node (MN) moves to a foreign network it obtains acare-of-address (COA) from the foreign agent (FA) that registers it with the home agent (HA)• COA is used by HA to forward packets destined to MN

SolicitationAdvertisement (FA,COA)

Register (HA)

Register


Home Agent (HA)

CorrespondentNode (CN)

Mobile Node (MN)

Packets to MN are picked up by the HAand tunneled to MN

Packets sent by MN godirectly to CN

• Triangle Routing in Mobile-IP


�� "��*

• Common issues in both paradigms– Location independent communication/addressing

– Location discovery/management

– Packet forwarding


� ��

• IP-Multicast– Single ‘logical’ multicast

group D-class address

– Senders do not knowwhere receivers are

– Receivers do not knowwhere senders are

• Mobile-IP– Permanent home address

– Temp care-of-address(es)

– Address mapping donethrough the home agent


� ��

• IP-Multicast– Membership location

– Done thru IGMP & routing(e.g., RP)

– Meet through the multicasttree

• Mobile-IP– Mobile node location

– Done thru home agent

– Meet thru registrationof new address


��4 ��

• IP-Multicast– Multicast forwarding

– Tunnel through themulticast tree (e.g., RP)

• Mobile-IP– Unicast forwarding

– Tunnel through homeagent


$��$!�+�&��+ �� "��*

• Instead of obtaining a new COA andregistering with the new foreign agent (andsubsequently with the home agent) and de-registering the old one

• Use the same logical multicast groupaddress and join/leave the group as youmove


(��"�� *��!��

CHCH: Correspondent host (sender)

Wireless link

Mobile Node


1 ��7��*�� +*��"�� CH

CH: Correspondent host (sender)

Wireless link

Mobile Node


$� �! �� ++

BS1 BS2


� ��)��

• Avoiding ‘triangle routing’ problem

• Avoiding the need for home/foreign agents

• Smooth hand-off using standard join/prune

• Using shortest path by source-specific trees


��.��(��

• Dynamics of the multicast tree during mobility

• Scoped multicast

• Crossing domain boundaries (inter-domainmovement)

• Join/Leave latency and smooth hand-off

• other issues…

Ahmed Helmy Assistant Professor of Computer Engineering ...afn48922/downs/wireless/Multicast_Tutorial.pdf · – DVMRP is a distance-vector routing protocol, – MOSPF is an extension

Documents