Appendix 1 - Multicasting, Multi-Unicasting, Port Numbers - Optelecom-nkf Siqura C-54 E-MC User Manual

Appendix 1: Multicasting, Multi-unicasting, Port numbers
APPENDIX 1 – MULTICASTING, MULTI-UNICASTING,
PORT NUMBERS
Multicasting
The C-54 E system can be used in a multicasting setting. The network switches and other
devices used must be carefully configured for, and capable of, handling
multicasting and its associated protocols (most notably IGMP v2). If not,
broadcasting will occur, which can put a very heavy load on the network. This is a
phenomenon inherent to multicasting and the facilities of network devices, not of the C-54 E
itself, although it is compounded by the density of the UDP streams used.
To define a multicast group, a source unit should be assigned a valid multicasting
('destination') TX stream address and the destination units should get this same address as
source. The group vanishes when the source is disabled, but the source will not
automatically be disabled when the last remaining destination is cancelled and will keep
transmitting at least towards the nearest switch. Additionally, it is possible to have the
multicast group units send unsolicited membership reports, keeping it alive even if only one -
any - unit of the group is still active.
Multi-unicasting
Alternatively, the C-54 E features 'multi-unicasting', i.e. sending out up to 2x3 independent
copies of video per video input and four data streams, as described in chapters 3, 4, 5 and 7.
If the bit rates selected are moderate, it may be more convenient to use this mechanism
instead of multicasting, even though the network gets more signal to carry from the encoder.
When such a destination is removed, the source also stops sending the corresponding
stream. If the input channel of a destination is disabled without disabling the source, source
transmission will be throttled, but not disabled (this behavior is selectable through the
FloodGuard settings discussed in section 7.7.2). The source downsizes the stream by sending
empty UDP packets until a wake-up call is received. The empty packets, of course, carry the
relevant IP/port information.
Port numbers
A valid UDP port number in a C-, S-, and V-series system is an unsigned 16-bit integer
between 1024 and 65536. Generally, you do not need to select other than the default
receiver port numbers as given in the MIB (Management Information Base). If you want to
change these receiver port numbers for some reason, use even numbers. A given receiver
port number N is associated with the port number N+1, through which control information is
returned to the source.
- 83 -