I would like to clarify one thing. Many people talked about issues or proposal based on the assumption network model in their brain. I don't think all that network model is the same. Why don't we clarify the target network model or parameters before leaving the start line?e.g. - how many node or MC ip address would be in one PIDWhat is your experience? There is some concern from our side, that the PID can not be easily used to separate/slit the address range, but itwould be extremely helpful if you could provide some insight from an operator.Operators we know typically use not more than a couple of PIDs for data services.
Actually we also used some PID for data service and user have not to move between PID in normal operation. This is unicast IP connectivity service. In other case, we studied to provide multicast data streaming service via satellite and it may have several PID and user will choose those PID in some way. At that time we didn't have mapping mechanism between Multicast IP address and PID using service information(SI) table. Therefore web based EPG like mechanism had been provided and the channel control order are sent to driver directory from the application. This is a sample mechanism and some other provider may employ similar system. So in terms of mapping mechanism of multicast ip address and PID is not necessary mechanism on L2. That means there are another way to map L2 and L3 address beside using DVB/MPEG tables. But in terms of unicast IP address I am not sure it is necessary or not.
Regards, Jun