Hello Venkat,
you have 8 branch routers connected via E1 links through an MPLS cloud to the central site.
The best type of multicast routing for your scenario is PIM sparse-mode.
However, if you are using a L3 MPLS VPN you have two ways:
-ask to your provider if it can cooperate with you providing a multicast MPLS VPN service (be prepared to pay additional fee for this)
- otherwise you need to build a mesh of GRE tunnels to carry your multicast packets inside unicast packets with GRE header.
You need to define a central router as the Rendezvous-point for all multicast groups.
This router provides a reference point for setting up multicast forwarding.
In case of SP provided MPLS multicast VPN it is the service provider that will provide you an RP. In that case follow their instructions.
On all routers do the following
enable ip multicast routing
ip multicast-routing
on each interface that can be source, destination or transit for multicast traffic enable pim sparse-mode
int type x/y
ip pim sparse-mode
( or ip pim sparse-dense-mode if the SP asks you to use autoRP see later)
the interfaces include the lan of the sources, client vlans in the central site, client vlans on branch sites, MPLS links on branch and on central site (see notes above this is enough if your MPLS service is L2 or the SP is giving you multicast support).
This list has to include backup /alternate interfaces as well.
RP information can be propagated:
a) with a static configuration with commands like
ip pim rp
b) using autoRP (may require ip pim sparse-dense-mode on all interfaces)
c) bootstrap
In you case a static configuration can be enough if SP assisted follow their instructions
for an introduction to multicast see from 12.4 multicast conf guide:
http://www.cisco.com/en/US/docs/ios/ipmulti/configuration/guide/imc_basic_cfg_ps6350_TSD_Products_Configuration_Guide_Chapter.html
Hope to help
Giuseppe
