Distributed digital rule processor for single system image on a clustered network and method

A distributed digital rule processor to create a single system image and execute digital rules on networked von Neumann processors comprising:

• a plurality of rule nets each having an ordered list of rules in a rule net storage, wherein each rule has inputs and outputs, the rule nets further comprising:
• a local controller coupled to the rule net storage to recognize broadcast rules;
• and a local bindery coupled to the local controller to store values, wherein the local bindery further comprises:
• a function state buffer to define which group of parallel instructions may be recognized next for execution by the individual rule net;
• a counter overflow unit for storing an object number and the maximum time allowed to a slave translator or rule net to complete the objects processing before a failure is reported;
• a plurality of rule parameter buffers for storing the parameters broadcast by the global controller with a rule;
• and
• a from function number buffer to store a number defining the rule net which sent a rule to a receiving rule net;
• a global controller having a global queue, a global bindery to store the state of global queue, and global data, to thereby distribute digital rules and control the flow of digital information between the rule nets;
• an Other Node Buffer to store return addresses of rules broadcast by the rule nets through the global controller and global bindery;
• and
• a plurality of slave translators each having their own memory, operating system, object handler, and an asynchronous communication connection with the global controller and global bindery wherein the slave translators process rules and data received from the rule nets and return that processed data to the rule nets.

A method for distributed digital rule processing to execute digital rules and create a single system image for a computer program on networked nodes having von Neumann processors, the method comprising:

• (a) compiling a plurality of rules;
• (b) storing the rules in a plurality of rule nets;
• (c) broadcasting a first digital rule from a global controller and a global queue;
• (d) accepting the broadcast rule if it is recognized by a specific rule net or slave translator;
• (e) updating a function state stored in one of a plurality of local binderies of the accepting rule net, wherein the function state and local bindery cannot be directly accessed by the rules in the rule net or the global controller;
• (f) processing the digital rule when the rule is accepted by one of the rule nets;
• (g) broadcasting a digital rule generated by the rule net as a result of the rule received by the rule net;
• (h) updating a function state stored in one of a plurality of local binderies of the broadcasting rule net, wherein the function state stores a function state value of the rule broadcast;
• (i) re-broadcasting a rule from a rule net which failed to complete on the node to which the rule was broadcast using the function state stored in the corresponding bindery, to thereby allow the program to continue execution without the loss of gray data;
• (j) storing the broadcast rule in the global queue to be broadcast by the global controller to other rule nets and slave translators;
• (k) accepting the broadcast rule when the rule is recognized by a specific rule net or slave translator;
• (l) processing rules received by a slave translator by calling slave application program interfaces (APIs) having slave von Neuman processors to process the digital rule and return the results to the calling rule net;
• and
• (m) repeating steps (d) through (l) until the processing of a distributed application is complete.