Method for distributed redundant execution of program modules

A method for distributed redundant execution of program modules in a distributed system which has at least two processors connected by transmission medium and in which only one version of a program module is present and stored in each one of at least two processors, said method comprising the steps of:

• asynchronously executing, in each of said at least two processors, said program module to produce result data, after receiving an input data message containing input data necessary for this execution;
• checking to conclusion, in each of said at least two processors which executed said program module, acceptability of the result data produced by said asynchronously executing said program module based only upon information within each of said at least two processors;
• sending, by each of said at least two processors which executed said program module and completed said checking with positive conclusion about acceptability, to at least one destination processor an information message which contains at least the result data produced by said execution of said program module through the transmission medium;
• and
• receiving, in said at least one destination processor, said information messages and selecting one of the received information messages as a new input data message by comparing the contents of the result data in said received information messages.

A method for distributed redundant execution of program modules in a distributed system which has at least two processors connected by transmission medium and in which only one version of a program module is present and stored in each one of at least two processors, said method comprising the steps of:

• asynchronously executing, in each of said at least two processors, said program module to produce result data, after receiving an input data message containing input data necessary for said execution;
• checking to conclusion, in each of said at least two processors which executed said program module, acceptability of the result data produced by said asynchronously executing said program module based only upon information within each of said at least two processors;
• sending, by each of said at least two processors which executed said program module and completed said checking with a positive conclusion about the acceptability, to at least one destination processor an information message which contains at least both the result data produced by said execution of said program module and said input data utilized during said program module execution through the transmission medium;
• and
• receiving, in said at least one destination processor, said information messages and selecting one of the received information messages as a new input data message based on both the result data and the input data contained in said received information messages.

A method for distributed redundant execution of program modules in a distributed system which has at least two processors connected by transmission medium and in which at least first and second different versions of a program module are stored in each one of at least two processors, said method comprising the steps of:

• asynchronously executing, in each of said at least two processors, any of said at least first and second different versions of a program module which is stored therein to produce result data, after receiving an input data message containing input data necessary for said asynchronous execution;
• sending, by each of said at least two processors which executed a version of the program module, to at least one destination processor an information message which contains at least both the result data produced by said asynchronous execution of said version and the information about said asynchronously executed version and about identity of said sending processor through the transmission medium;
• receiving, in said at least one destination processor, said information messages and selecting an acceptable one of the received information messages as a new input data message;
• and
• sending, from each destination processor which selected one of the received information messages, a second message which identifies which of said at least two processors produced said information messages that were discovered to be unacceptable during said selecting step, to the processors that produced said information messages.

A method for distributed redundant execution of program modules in a distributed system which has at least two processors connected by transmission medium and in which at least first and second different versions of a program module are stored in each one of at least two processors, said method comprising the steps of:

• asynchronously executing, in each of said at least two processors, any one of said at least first and second different versions of a program module which is stored therein to produce result data, after receiving an input data message containing input data necessary for said asynchronous execution;
• sending, by each of said at least two processors, to at least one destination processor an information message which contains at least:
• the result data produced by said execution of said one of said at least first and second different versions;
• said input data utilized during said execution of said at least first and second different versions;
• and information about said executed version through the transmission medium:
• receiving, in said at least one destination processor, said information messages and selecting one of the information messages as a new input data message based on the result data and the input data contained in said information messages;
• and
• sending, from said each destination processor which selected one of the received information messages, a second message which identifies which of said at least two processors produced said information messages that were discovered to be unacceptable during said selecting step to the processor that produced said information message.

A method for distributed redundant execution of program modules in a distributed system which has at least two processors connected by transmission medium and in which at least first and second different versions of a program module are stored in each of at least two processors, said method comprising the steps of:

• asynchronously executing, in each of said at least two processors, any one of said at least first and second different versions of a program module to produce result data, after receiving an input data message containing input data necessary for said asynchronous execution;
• checking to conclusion, in each of said at least two processors which executed said any one of said at least first and second different versions of the program module, acceptability of the result data produced by said execution based only upon information within each of said at least two processors;
• sending, by each of at least two processors which executed one of said at least first and second different versions of the program module, to at least one destination processor an information message which contains at least:
• the result data produced by said execution of said any one version and
• the information about said asynchronously executed version and about identity of said sending processor through the transmission medium;
• and
• receiving, in said at least one destination processor, said information messages and selecting one of the information messages as a new input data message.

A method for distributed redundant execution of program modules in a distributed system which has at least two processors connected by transmission medium and in which at least first and second different versions of a program module are stored in each of at least two processors, said method comprising the steps of:

• asynchronously executing, in each of said at least two processors, any of said at least first and second different versions of a program module which is stored therein to produce result data, after receiving a message containing input data necessary for said asynchronous execution;
• checking to conclusion, in each of said at least two processors which executed said one of said at least first and second different versions of the program module, acceptability of the result data produced by said asynchronously executing said one version of said program module based only upon information within each of said at least two processors;
• sending, by each of at least two processors which executed one of said at least first and second different versions of the program module, to at least one destination processor an information message which contains at least:
• the result data produced by said asynchronously executing said one version of said program module;
• said input data utilized during said one version execution;
• and
• information about said executed version and about identity of said sending processor through the transmission medium;
• and
• receiving, in said at least one destination processor, said information messages and selecting one of the information messages as a new input data message based on the result data and the input data contained in said received information messages.