Time-triggered process for the periodic fault-tolerant transmission of real-time data in a distributed computer system
The invention relates to a method for periodic transmission of real time data in a computer system, particularly a distributed computer system, which computer system is comprised of node computers ( 201 - 208 ), particularly an appreciable number of node computers ( 201 - 208 ), and distributor units ( 211 - 215 ), particularly an appreciable number of distributor units ( 211 - 215 ), wherein the node computers ( 201 - 208 ) and the distributor units ( 211 - 215 ) have access to a global time, and wherein real time data are transmitted by means of time-triggered real time messages, wherein selected distributor units ( 212, 213, 214, 215 ) form a central structure of distributor units; and wherein during a periodic communication round (PCR), in the error-free case, at least two copies of each real time message to be sent are transmitted via at least two independent routes through the central structure, by executing a satisfying or an optimal time plan, from a start distributor unit in the central structure to a target distributor unit in the central structure; such that in the error case, error handling can begin immediately after an error detection time point contained in the time plan, preferably in the optimal time plan.
1. A time-triggered fault-tolerant method for periodic transmission of real time data in a distributed computer system, which distributed computer system is comprised of a plurality of node computers ( 201 - 208 ) and a plurality of distributor units ( 211 - 215 ), wherein the node computers ( 201 - 208 ) and the distributor units ( 211 - 215 ) have access to a global time, and wherein real time data are transmitted by time-triggered real time messages, the method comprising:
selecting distributor units ( 212 , 213 , 214 , 215 ) to form a central structure of distributor units; and
transmitting during a periodic communication round (PCR), in an error-free case, at least two copies of each of the time-triggered real time messages to be sent via at least two independent routes through the central structure, by executing a satisfactory or optimal time plan, from a start distributor unit in the central structure to a target distributor unit in the central structure such that in an error case, error handling begins immediately after an error detection time point contained in the satisfactory or optimal time plan.
2. The method according to claim 1 , wherein after an end of a periodic communication round, those distributor units which have participated in this periodic communication round send a copy of the time-triggered real time messages which have been received during this periodic communication round, which time-triggered real time messages are correct, to the node computers associated with the respective distributor unit.
3. The method according to claim 1 , wherein after an end of a periodic communication round, the distributor units report errors which have been detected during the periodic communication round to a monitor node computer.
4. The method according to claim 1 , wherein the transmission of time-triggered real time messages in the central structure of distributor units is carried out with use of TTEthernet protocol.
5. The method according to claim 1 , wherein the start distributor unit transmits the real time data in a PCR to each distributor unit in the central structure.
6. The method according to claim 1 , wherein all distributor units in the central structure function as respective start distributor units, which in a PCR transmit time-triggered real time messages to target distributor units comprising all other selected distributor units in the central structure.
7. A distributed computer system, comprising:
a plurality of node computers ( 201 - 208 ); and
a plurality of distributor units ( 211 - 215 ), wherein the node computers ( 201 - 208 ) and the distributor units ( 211 - 215 ) have access to a global time, and wherein real time data are transmitted by time-triggered real time messages,
wherein selected distributor units ( 212 , 213 , 214 , 215 ) form a central structure of distributor units, and
wherein during a periodic communication round (PCR), in the error-free case, at least two copies of each time-triggered real time message to be sent are transmitted via at least two independent routes through the central structure, by executing a satisfactory or optimal time plan, from a start distributor unit in the central structure to a target distributor unit in the central structure such that in the error case, error handling begins immediately after an error detection time point contained in the satisfactory or optimal time plan.
8. The system according to claim 7 , wherein after an end of a periodic communication round, those distributor units which have participated in this periodic communication round send a copy of the time-triggered real time messages which have been received during this periodic communication round, which time-triggered real time messages are correct, to the node computers associated with the respective distributor unit.
9. The system according to claim 7 , wherein after an end of a periodic communication round, the distributor units are configured to report errors which have been detected during the periodic communication round to a monitor node computer.
10. The system according to claim 7 , wherein the transmission of time-triggered real time messages in the central structure of distributor units is configured to be carried out with use of TTEthernet protocol.
11. The system according to claim 7 , wherein the start distributor unit is configured to transmit the real time data in a PCR to each distributor unit in the central structure.
12. The system according to claim 7 , wherein ail distributor units in the central structure are configured to function as respective start distributor units, which in a PCR transmit time-triggered real time messages to target distributor units comprising all other selected distributor units in the central structure.
13. The system according to claim 7 , wherein a node computer ( 208 ) with two or more selected distributor units ( 212 , 215 ) is connected to the central structure of distributor units.