Re-Activating Senders & Receivers
If an explicit activation is performed against a Sender or Receiver, the API must request a re-application of settings to the underlying Sender or Receiver implementation whether the setting have changed or not. For example in the case of multicast Receivers it is suggested that this involves an explicit IGMP leave and join. For a Sender this may involve stopping and re-starting the stream.
Transport Files & Caching
It is strongly recommended that the following caching headers are included via the /transportfile endpoint (or whatever this endpoint redirects to).
This is important to ensure that connection management clients do not cache the contents of transport files which are liable to change.
In environments where multiple clients may be operating against a single Connection Management API, it is possible that staging of parameters may result in conflicts. There is intentionally no mechanism to prevent this in the API, however clients should examine the results of HTTP PATCH operations which will return the full complement of current settings, allowing the client to confirm that only the changes it had requested have been made.
Where a server implementation supports concurrent application of settings changes to underlying Senders and Receivers, it may choose to perform ‘bulk’ resource operations in a parallel fashion internally. This is an implementation decision and is not a requirement of this specification.
If a ‘bulk’ request includes multiple sets of parameters for the same Sender or Receiver ID the behaviour is defined by the implementation. In order to maximise interoperability clients are encouraged not to include the same Sender or Receiver ID multiple times in the same ‘bulk’ request.
IS-05 provides a mechanism whereby the activation of staged transport parameters can be performed at a particular TAI time, or after a given interval. The primary use case for this behaviour is to allow the synchronisation of large salvo operations, where multiple pieces of equipment should carry out a change in transport parameters simultaneously. A controller may stage parameters with multiple IS-05 APIs with the same activation timestamp, and know that activation will occur at the same time on all devices regardless of the latency between the controller and the device. It is not intended to act as a mechanism for scheduling activations far in the future.
In the event of an error occurring between scheduling an activation and the activation time, the IS-05 transport parameters should reflect the current state of the device. For example, if a sender is no longer sending data at all it should reflect this by setting
master_enable to false.
API clients should check back with the API after the expected activation time to ensure activations have completed successfully, and that the device is in the expected state. Note that a client may normally expect to see a change in the IS-04 version timestamp upon a successful update of transport parameters, and should monitor for this version timestamp update via the IS-04 web-socket where IS-05 is being used along-side IS-04.