Method for switching between redundant buses in a distributed processing system

摘要

When an error is detected during transmission of data in a distributed processing system having redundant unidirectional circulating buses, an error procedure is entered in each of the processors detecting the error. In the procedure, a testing processor transmits a test message to itself, first on the active bus and then on the inactive bus if the test message is unsuccessfully transmitted on the active bus. If the test message is successfully transmitted on the inactive bus, the active and inactive buses are switched and a fault is declared on the previously active bus. Other processors receiving the fault signal perform the same tests themselves and switch to the previously inactive bus if their test of the active bus is also unsuccessful. If the test of the active bus was successful, a second test is performed and if that is also successful, operation is resumed on the active bus. If the tests on both buses are unsuccessful operation also resumes on the originally active bus. This avoids repeated switching between buses.