You are on page 1of 3

Multicast Server = Source

Multicast Client = Destination

IGMP protocol is used by client to join a multicast group. IGMP protocol is turn
ed ON as soon as we define PIM mode.
"ip igmp join-group" command is used for lab purposed to emulate the router inte
rface as client. it is used for testing where the router is respond back to ping
to the multicast group.
"ip igmp static-group" does not respond back to pings to the multicast group. it
is seldomly used. one example of its use is when you dont want multicast client
s to send multicast join group messages and would rather flood out the multicast
traffic on the interface.
another use can be in setting up span where you put this command on the span int
erface.
so we are using ip igmp join-group to emulate multicast client. how do we emulat
e multicast server. a servere is the end that sends the traffic. we can use ping
traffic from the source to emulate it as the server.
so from source to destination ping, source ip is unicast and destination ip is m
ulticast
ping response from destination, source and destination are both unicast ip addre
sses
in real scenario, source ( server ) send the multicast traffic and destination (
clients ) receive it.
each router run RPF check along the path.

MULTICAST TRAFFIC FLOW:


-------------------------
=>server generates traffic.
=>server doesnt know if the multicast clients are local or remote and it dont ca
re. it just streams the traffic out.
=>traffic reaches the first router on the server side
=>router mroute routing table is empty at first
=>but we now have s,g ( source , group-address ) and/or *,g entry => source is s
erver ip address and group-address is the multicast group-address.
on the client side router, we have client talking to the router via igmp so we w
ill have *,g entry and/or s,g entry
=>routers along the path will send and receive traffic on the interfaces that ha
s pim mode configured and will have entries in the mroute table.
DENSE MODE:
-------------
push technology
dont have to do much. server push traffic to the client and it goes along the pa
th finally to the destination.
SPARSE MODE:
------------
explicit join ( no traffic unless you ask for it )
Rendezvous Point ( RP ) is where Routers that know about clients that want the t
raffic and Routers that know about multicast sources.
rp can be configured
MANUALLY: via ip pim rp-address command that goes on every router along the path
AUTOMATICALLY: via auto-rp or bootstrap router (aka pimv2)
HOW RP DISCOVER SOURCE ( server ):
----------------------------------
=>source S1 sends traffic to group G1
=>PIM DR on LAN segment ( i think its the first hop router but we dont have to w
orry about it as we dont have to specify who is the DR ) hears (S1,G1) traffic a
nd sends unicast register message ( encapsulates the whole multicast packet in u
nicast with his unicast address as source and RP unicast address as destination
) to RP
( "debug ip rp" on DR and RP shows this )
=>RP now knows that (S1,G1) is sending and replies to DR with register stop
(i know the source exists, stop sending me messages )
=>DR will periodically refresh register messages
=> routers in transit up along the path from DR to RP will never have s,g or *,g
entry since traffic from DR to RP is sent via unicast
HOW RP DISCOVER DESTINATION ( client ):
---------------------------------------
=> destination ( client ) sends igmp join for ( *,G1 ) onto LAN at destination i
p address of multicast group.
=> PIM DR on LAN segment hears IGMP join for ( *.G1 ) and sends PIM join for ( *
,G1 ) up reverse path tree towards RP
( "debug ip pim" on DR,RP, and transit path will show that )
=> PIM Routers in transit path install ( *,G1 ) entry in multicast routing table
=> RP now knows that a host wants ( *,G1 )

COMMANDS:
-----------
Multicast is something we do after we have fully completed unicast routing.Any c
hanges after multicast config is setup in the unicast routing can cause RPF chec
k failure. so we gotta be careful about that/
(config)#ip multicast-routing distributed //distributed keyword used in 35
60 switches // enables
multicasting
(config-if)#ip pim dense-mode //command goes on all interface that are taki
ng part in multicasting.
//it is the only command used in addition to
enabling multicasting for //dense mode
(config-if)#ip pim sparse-mode // command goes on all interfaces that are t
aking part in multicasting
//use this command even on the RP router inte
rfaces including loopback //if in use
(config)#ip pim rp-address x.x.x.x // static rp assignment
//use this command on all the routers in the
path
// x.x.x.x can be the loopback of the RP rout
er
//use this command even on the RP router
show ip pim neighbor
show ip mroute
sho ip pim mapping

AUTO-RP:
========
defines 2 roles
- candidate RP
uses group 224.0.1.39 to advertise itself ( its unicast address or somet
hing ) to the mapping agent.
other routers ( the ones in the path ) dont listen to group 224.0.1.39
-> ip pim send-rp-announce
- mapping agent
uses group 224.0.1.40 to advertise RP info to other routers.
-> ip pim send-rp-discovery
RPF check on client routers is done on auto-RP announcements

You might also like