[Xorp-users] IGMP/PIM question ?

[Pavlin Radoslavov]

> > Also, they won't eliminate the PIM Register encapsulation issue.
> > Though, if you always switch to the shortest path (to the sender)
> > after the first data packet, then the PIM Register duplicated packet
> > will happen only when a new sender becomes active (until the SPT
> > switch to that sender is completed).
> 
> If you can live with the initial data duplicates as described above,
> then you need to hack only the PIM-SM DR election (and configure
> your PIM-SM routers to perform the SPT switch after the first data
> packet).

On a third thought, you could hack the PIM-SM implementation to
ignore WHOLEPKT kernel upcalls that indicate the sender address
doesn't belong to the router. This will automatically prevent the
PIM Register encapsulation. Note that this will work for Linux, but
in case of *BSD you also need to use the
--disable-advanced-multicast-api configure switch to disable the PIM
Register encapsulation inside the kernel (the Linux kernel doesn't
implement the Register encapsulation inside the kernel).

To achieve that you need to modify
PimMrt::signal_message_wholepkt_recv() inside pim_mrt_mfc.cc.
With such modification you don't need to worry about the SPT switch
threshold.

Pavlin