Provider Aborts

A CPDLC Provider Abort (PA) is asserted locally at the ATN application layer, and gives rise to a CPDLC-provider-abort indication. In principle, a PA may be asserted both by the ground and the air side, but since most PAs are associated with a loss of end-to-end communication, only those asserted by the ground application are normally recognised by the ANSP and hence the CRO LISAT database.

The ATN CPDLC application SARPs specify a ProviderAbortReason code which identifies the cause of the PA. Generally (but not exclusively), a PA arises when a sustained loss of end-to-end connectivity lasts for longer than 6 minutes (the transport layer inactivity timer value ), causing the underlying transport level connection to fail. This is typically seen at the transport layer by a Disconnect Request with a Reason code not set to Normal (0x80). Such loss of connectivity may be due to failure of the air-ground link, or may also be associated with a malfunction at the ATN Network layer (e.g. ATN router problems). Rarely, a PA may also arise from a malfunction or protocol error at the transport or application layers, while the communication service itself remains functional. Some examples of the latter scenario have been observed during the LINK 2000+ Pioneer phase, but the causes have since been fixed, and accordingly such cases are now very uncommon.

Before the transport layer inactivity timer triggers a PA, the air or ground ATN End System (or both), may have faced transport level retransmissions. In order for the transport level retransmission process not to cause the transport level connection to fail before the Inactivity timer expires, it is important that the transport retransmission count is correctly configured on both the air and ground End Systems. The DLS EUROCONTROL Specification recommended value for the retransmission count is 7; an incorrectly set lower value would lead an End System to face a PA prematurely compared to the remote End System resulting in desynchronisation of the air and ground systems.

It should also be noted that the ACSPs do not routinely inspect the content of transport or application layer PDUs, and so will not necessarily be aware of individual PA events, although they may detect other evidence of failure of the underlying communication service giving rise to a PA.