BGP Multicast

Multicast BGP
Some Cases don’t want uni-cast traffic to route the same direction as multicast traffic, multicast BGP is basically a way to change the Reverse Path Forwarding Check without having to do static multicast routes everywhere

For multicast transit over the internet all hops must run multicast

• What if the RPF check for a multicast source is via a unicast only peer?
• Multicast BGP solves this by separating unicast RPF and multicast RPF
• Multicast BGP advertises source networks for purpose of RPF check
• By the way, Multicast BGP does not replace PIM
• Multicast BGP preferred over uni-cast protocols for multicast RPF check
• Like a static multicast route but dynamic
•     Doesn’t require a separate routing protocol, only BGP extensions

       How Multicast BGP Works

•       BGP peers negotiate Multicast Address Family during capabilities exchange
•       Peers advertises NLRI under Multicast Address Family
•       Network statement just like uni-cast
•       All normal BGP rules apply
•        When multicast traffic is received, MBGP learned routes are preferred over uni-cast

        Configuration

       Note - *** eBGP (which is AS- 20 lowest AS number) might want to use. If you want to use iBGP  (  which is AS-200), route will prefer IGP which is lower AS than iBGP and RPF check will fail. This point, you might want to change AS number lager than iBGP  **** 

•        Router (config) # router bgp 200
•        Router (config-router) # neighbor 10.35.35.5 remote-as 1
•        Router (config-router) # address-family ipv4 unicast
•        Router (config-router) # neighbor 10.35.35.5 activate
•        Router (config-router) # address-family ipv4 multicast
•        Router (config-router) # neighbor 10.35.35.5 activate
•        Router (config-router-af) # network 10.45.45.0 mask 255.255.255.0

       BGP Multicast LAB

     R3 CONFIGURATION

       R3(config-router-af)#do sh run | section router bgp

       router bgp 2

       bgp log-neighbor-changes 

       neighbor 10.35.35.5 remote-as 1 

       !

       address-family ipv4 

       neighbor 10.35.35.5 activate

       exit-address-family

       ! 

       address-family ipv4 multicast

       network 10.12.12.0 mask 255.255.255.0

       network 10.45.45.0 mask 255.255.255.0

       neighbor 10.35.35.5 activate

       exit-address-family

       R3(config-router-af)#

    

       R5 CONFIGURATION

       R5#show run | section router bgp

       router bgp 1 

       bgp log-neighbor-changes

       neighbor 10.35.35.3 remote-as 2

       !

       address-family ipv4

       neighbor 10.35.35.3 activate 

       exit-address-family 

       ! 

       address-family ipv4 multicast

       network 2.2.2.2 mask 255.255.255.255 

       network 10.45.45.0 mask 255.255.255.0 

       neighbor 10.35.35.3 activate 

       exit-address-family

   ROUTER 5 RPF CHECK ROUTE

•      Even though R5 unicast route through R3 but RPF route go through R4.

      Check route from R5 to R1 multicast and uni-cast routing   

     REFERENCE

         https://ine.com/
   https://www.cisco.com/c/en/us/td/docs/ios/12_2/ip/configuration/guide/fipr_c/1cfmbgp.html
   https://www.cisco.com/c/en/us/td/docs/ios/12_2/iproute/command/reference/fiprrp_r/1rfmbgp.pdf 

  

BIDIR (BIDIRECTIONAL PIM)

BIDIR ( Bidirectional PIM)

Use cases for this bidirectional is when your application is many to many where you are both the sender and the receiver and want to cut down the number of entries in the routing table 

• We can use Bidir when we have lots of senders that are also receivers in the Multicast Network
•          Traditional sparse mode forms two trees

o   Unidirectional SPT from source to RP

o   Unidirectional shared tree from RP to receivers

•           Results in (*,G) and (S,G) entries in control plane

o   For many to many multicast application, doesn’t scale well

• Bidirectional PIM solves by only allowing the shared tree (*,G) and never a SPT(S,G)

  

 How Bidirectional PIM works

•      Defie an RP and group rnage as bidirectional

o   Stops formation of (S,G) for that range

•       Build single (*,G) tree towarfs RP

o   Traffic flows upstream from source to RP

o   Traffic flows downstream from RP to receivers

•       Removes PIM Register process

o   Implies that traffic from sources always flows to the RP          

•        Uses designated forwarder (DF) for loop prevention
•         One  DF is elected per PIM segment

o   Similar to assert, lowest metric to the RP wins

o   Highest IP address in tie

•        Only DF can forward traffic upstream towards RP
•        All other interfaces in OIL are downstream facing
•         Removes the need for RPF check

o   Due to this all routers must agree on Bidir or loops can occur

 

  Bidir Configuration

•        Router (config) # ip multicast-routing
•         Router (config)#  ip pim bidir-enable
•          Router (config)#  ip pim rp-address x.x.x.x bdir

  SHOW command

•          Router # show ip mroute
     ·  Router # show ip pm rp mapping
     ·  Router # show run | inc bidir
     ·   Router # show ip mfib count

   LAB CONFIGURATION

  Router 2 configuration 

      This lab router 2 is RP bi-direcitonal 

   

   Checking mroute for routers

        After ip igmp join group 224.1.1.1 

   REFERENCE

•         https://ine.com
•         https://www.cisco.com/c/en/us/td/docs/ios/12_0s/feature/guide/fsbidir.html
•         https://www.cisco.com/c/dam/en/us/products/collateral/ios-nx-os-software/multicast-enterprise/prod_white_paper0900aecd80310db2.pdf