As IP multicast is being more widely deployed and used, the
existing multicast routing algorithms have demonstrated several
limitations which make them unsuitable for deployment globally
or among multiple provider domains.  Protocols like DVMRP and
PIM Dense Mode that rely on broadcasting and pruning leave
state in parts of the network that are not on the multicast
delivery tree.  Protocols like CBT and PIM Sparse Mode use a
centralized resource to learn of multicast sources.  Service
providers are reluctant to maintain state for multicast groups
that have no receivers in their domain or use a centralized
resource in another domain that they cannot control.

BGMP is a scalable multicast routing protocol which addresses
these problems.  Like CBT and PIM Sparse Mode, BGMP chooses a
global root for a delivery tree.  However, the root is a domain,
not a single router, so if there is any path available to the
domain connectivity can be maintained.  BGMP builds a bidirectional,
shared tree of domains.  Similarly to the unicast EGP/IGP split,
BGMP is used as the inter-domain or external protocol, while
domains can run any multicast IGP internally (such as CBT or
PIM Sparse Mode), and can build source-specific shortest-path
distribution branches to supplant the shared tree where needed.

The BGMP working group is chartered to complete the protocol
specification and follow it through the Internet standards
track.  It will also help to design a transition mechanism
from MSDP (the Multicast Source Distribution Protocol, an
interim interdomain solution that is unlikely to scale for
the long term) to Internet-wide BGMP.