A British-Canadian company was set up to develop a Care-Alarm System which could interface to existing equipment provided by other manufacturers, most of whom used proprietary protocols.
The hardware in the response centre consisted of a number of networked Unix workstations which served as operator consoles. These were networked to a PC running Unix which included an AT&T DSP processor card to handle the telephony.
RTD developed all the Unix software for the PC which provided control and management functions for the telephony subsystem. One of the major tasks of the software was to ensure that, in the event of a catastrophic error, the system recovered quickly without losing calls.
System design considerations required that a datagram protocol be used in communicating with the workstation, yet error detection and correction was essential. RTD designed an efficient error-correcting protocol which met all the system requirements in terms of speed and reliability.
The system was required to work even in the event of a network failure. RTD designed a module which ran on the PC itself and provided all the call control facilities of the workstations but with a much simpler user interface. This system came into operation automatically in the event of a network failure. The only feature which was not provided was database access as the database resided on the workstations. Nevertheless, the system ensured that the database was correctly updated when the network recovered.