Hi ambi,
In my following answers, I assume that we have a simple Layer2 switch that has no specific support for IGMP Snooping or MLD Snooping - let's just assume it is a plain simple switch like the one installed into a common home DSL router. If you want to know how the switches with IGMP Snooping or MLD Snooping react then please say so - keep in mind that things are more complex on those switches.
How does a switch treat a multicast packet (non igmp) when there is no source sending traffic for that multicast group.
I do not understand this statement. If there is a multicast packet then someone must have sent it - and that device is the source of that multicast packet, right?
In any case, a simple basic Layer2 switch will flood any multicast packet out all ports (in the same VLAN if the switch supports VLANs) except the incoming port.
how does an L2 switch treat a ipv6 MLD packet (which is multicast) when neither ipv6 or MLD is configured on the routers
A simple basic Layer2 switch will flood such MLD packet out all ports (in the same VLAN) except the incoming port.
how are the switches coded to treat well known multicast mac address like 33:33:xx:xx:xx:xx etc
They are not coded for such multicast MAC addresses at all. It is completely sufficient that they operate using basic bridging/switching rules that have existed since the first bridges and switches came to existence: If a switch does not know the destination MAC address, it will flood the frame out all ports except the incoming port. Because multicast MAC addresses such as 33:33:xx:xx:xx:xx are never used as source MAC addresses and thus can never be learned by a switch, they will always be treated as unknown destinations for which the traffic is always flooded. You do not need any specific support on switches for basic multicast delivery.
Best regards,
Peter
