The Bendix Corporation

Omur Tasar

Arliss E Whiteside

Morris D Freedman

Alexander M Rothschild

A Fault-Tolerant Multi-Computer System for control applications is disclosed. The system has a plurality of Computers (10a-10n), each having an assigned set of tasks which it is capable of executing. No one Computer in the system acts as a master and no one Computer executes all of the tasks. Communication between the Computers is by individual communication links (16, 18, 20) over which each Computer sends information directly to all other Computers in the system. Each Computer comprises an Applications Computer (100) and an Operations Controller (200). The Operations Controller receives messages over the communication links and selects, from the assigned tasks, the tasks to be performed by the associated Applications Computer. Each Operations Controller includes a fault handler which checks the messages received from the other Computers. The fault handlers send and receive error messages, over the communication links, to assist in the identification of a faulty Computer. Subsequent messages from the Computers deemed to be faulty are ignored, and the tasks assigned to the faulty Computer are executed by alternate Computers in the system.

Fault-tolerant multi-computer system

A task communicator for each computer in a multiple computer system is disclosed. The task communicator provides communication of data values between cooperating tasks executed by different computers. The task communicator comprises a data values table (804) storing the current data values required for the execution of each task assigned to its computer. A store data value module (802) which stores the data values received from the other computers in the data values table. A task dispatcher (806) which copies the data values required for the execution of a selected task into a task input table (810). A release task module (814), which releases for execution the identified task and data values stored in the task input table. A task output table (812) storing the values of the data resulting from the execution of each task by its own computer. A task results message sender (818) which sends to all of the computers in the system messages containing the new data values stored in the task output table by the executed task.

Task communicator for multiple computer system

An operations controller for each computer in a multiple computer system is disclosed. Each operations controller controls the operations of its associated computer, so that all of the computers cooperate to perform system functions in a fault-tolerant manner. Each operations controller comprises a fault handler (204), a scheduler (206), a task communicator (208), plus a transmitter (212) and requisite receivers (202) which receive and send messages to all the other computers in the system. The fault handler (204) checks each message received and decides which computers are operating correctly and which are faulty. A scheduler (206) selects each task its own computer will execute, from the tasks assigned to its own computer. A task communicator (208) assembles the data values required for the execution of the selected task, and forwards this data to the computer for execution. The operations controller sends messages to all of the other computers in the system, informing them of which computer it deems to be faulty, the tasks it selects to execute, when it starts and completes the execution of each selected task, and the data variable values produced by the execution of each task.

Operations controller for a fault-tolerant multiple computer system

A fault handler for each computer in a multiple computer system is disclosed, which excludes faulty computers from participating in the operation of the multiple computer system. The fault handler comprises one or more message checkers (216, 218, 220, 222 and 224) and a fault tolerator (228). In the disclosed embodiment, the fault handler further includes a synchronizer (226) for synchronizing the operation of the associated computer with the operation of the other computers in the system. The checker modules check each message received from the other computers, and forward the messages to the fault tolerator. From error messages received from the other computers and errors detected by its own checker modules, the fault tolerator (228) decides which computers are faulty and discards the messages received from those computers. Only messages received from non-faulty computers are passed on for further processing. When an error is detected in a message, the fault tolerator sends a message to all of the other computers identifying the computer which sent the message.

Growing community of inventors

Harvard, MA, United States of America

Omur Tasar