System and method for performing distributed asynchronous calculations in a networked environment

1. A method for performing calculations in a distributed networked environment, comprising:

updating a source element on a first computer;

sending a calculate update signal which is a message communicated through a message queuing system in response to the updating, specifically directed to a target calculated element for updating on a target element computer that may be the first computer or a second computer connected to the first computer that uses the source element in a calculation;

obtaining the updated source element by a process on the target element computer associated with the target calculated element in response to the sending;

performing a calculation to update the target calculated element using the obtained updated source element;

dispatching, by a dispatcher, a first and second update message directed to a first target element;

bundling the first and second update message into a bundled message;

assigning a bundled message identifier to the bundled message;

sending the bundled message to a shard processor that handles the bundled message;

applying each message in the bundled message to the first target element; and

acknowledging, by the shard processor to the dispatcher, application of each bundled message.

2. The method according to claim 1 , further comprising:

determining, by the dispatcher, a state of completeness if all bundled messages have been acknowledged.

3. The method according to claim 1 , further comprising:

dispatching, by the dispatcher, a third and fourth update message directed to a second target element;

bundling the third and fourth update message into a second bundled message;

assigning a second bundled message identifier to the second bundled message; and

sending the second bundled message to the shard processor that handled the bundled message.

4. The method according to claim 3 , further comprising:

assigning a common graph identifier to each of the first target element and the second target element that are related to each other;

splitting a graph containing the first target element and the second target element after the assigning; and

maintaining the common graph identifier in each of the first target element and the second target element after the split.

5. The method according to claim 1 , further comprising:

blocking the performing of the calculation until a further criterion is met.

6. A method for performing calculations in a distributed networked environment, comprising:

updating a source element on a first computer;

sending a calculate update signal which is a message communicated through a message queuing system in response to the updating, specifically directed to a target calculated element for updating on a target element computer that may be the first computer or a second computer connected to the first computer that uses the source element in a calculation;

obtaining the updated source element by a process on the target element computer associated with the target calculated element in response to the sending; and

performing a calculation to update the target calculated element using the obtained updated source element

the using of the output list for one of the source elements comprises:

a pending list of target elements that have not indicated having seen a current version of the source element; and

a completed list of target elements that have indicated having seen a current version of the source element;

the method further comprising:

moving one of the target elements from the pending list to the completed list when that target element indicates it has seen the current version of the source element; and

storing the other elements that have seen the element at its current version.

7. A method for performing calculations in a distributed networked environment, comprising:

updating a source element on a first computer;

sending a calculate update signal which is a message communicated through a message queuing system in response to the updating, specifically directed to a target calculated element for updating on a target element computer that may be the first computer or a second computer connected to the first computer that uses the source element in a calculation;

obtaining the updated source element by a process on the target element computer associated with the target calculated element in response to the sending;

performing a calculation to update the target calculated element using the obtained updated source element;

defining a graph as a plurality of elements;

creating a graph table comprising a plurality of records, wherein each record corresponds to a table element that is one of the plurality of elements, each record comprising:

an element number;

a revision number;

an expected count number;

a pending element identifier list that is a list of elements expected to reference this element but that have not yet seen the element at its current revision; and

a complete element identifier list that is a list of elements which have seen this element at its current version;

updating a record in response to an element calculate signal;

wherein:

each table element is designated as being in a state of complete, incomplete, or unseen;

an element in the complete state is one in which the element is seen and the pending list is empty;

an element in the incomplete state is one in which the pending list is not empty; and

an element in the unseen state is one for which no signal has been received, but that is used by an element that has been seen.

8. The method of claim 7 , further comprising:

determining that the graph is complete when all table elements are in a state of complete.

9. The method of claim 7 , further comprising:

reconstructing calculations based on a stored graph table in response to a determination of a system error in which update signals are lost.