Maintaining a relationship between two different items of data

Abstract

Data is stored persistently. At least two different items of the data are stored in two different non-conflicting regions or two different physical clusters. A relationship is maintained between the two different items of data. The relationship enables a process to reach any one of the data items from the other data item. Consistency of the relationship is maintained notwithstanding updates of either or both of the items.

1 - 23 . (canceled) 
     
     
         24 . A method comprising
 storing data items of a database persistently in a first physical cluster of a storage medium,   storing in other physical clusters of a storage medium data items that are replicas of the data items of the database that are stored in the first physical cluster,   partitioning the stored data items and the stored replicas of the data items into partitions that enable an algorithm that uses the data items to be decomposed into sub-algorithms that can be processed concurrently, and   maintaining consistency between the data items and the replicas of the data items as the data items or the replicas of the data items are updated.   
     
     
         25 . The method of  claim 24  in which each replica contains only data that is required by one of the sub-algorithms. 
     
     
         26 . The method of  claim 24  in which the selection of data for each of the replicas is a selection of items of data, or of elements of items, or of both. 
     
     
         27 . The method of  claim 24  in which the replicas are provided persistently as objects to an object-oriented application. 
     
     
         28 . The method of  claim 27  in which an object relational broker provides persistent storage of objects for an object-oriented application. 
     
     
         29 . The method of  claim 24  in which the data is stored in an object database. 
     
     
         30 . The method of  claim 24  in which the data is stored in a relational database with object-oriented extensions. 
     
     
         31 . The method of  claim 24  in which consistency between the data items and the replicas is maintained by concurrent processing on at least two different processors. 
     
     
         32 . The method of  claim 31  in which the two different processors do not use shared memory for the concurrent processing. 
     
     
         33 . The method of  claim 24  in which, while one of the replicas is being updated, the related portions of the data may be concurrently updated. 
     
     
         34 . The method of  claim 24  in which the data items in the replicas can be concurrently updated by at least two different processors. 
     
     
         35 . The method of  claim 24  in which the portions of the data included in each of the replicas are specified by a user. 
     
     
         36 . The method of  claim 24  in which the portions of the data included in each of the replicas are selected. 
     
     
         37 . The method of  claim 36  in which the selection applies to items of the data. 
     
     
         38 . The method of  claim 36  in which the selection applies to elements of items of the data. 
     
     
         39 . The method of  claim 36  in which the selection applies to items of data and to elements of items of the data. 
     
     
         40 . The method of  claim 36  in which the selection is based on different criteria for different replicated sets. 
     
     
         41 . The method of  claim 24  in which the replicas are physically clustered in storage. 
     
     
         42 . The method of  claim 24  in which the replicas are respectively physically clustered. 
     
     
         43 . The method of  claim 24  in which the replicas are associated with indexes. 
     
     
         44 . The method of  claim 43  in which the indexes include replicated data for use by the algorithm. 
     
     
         45 . The method of  claim 44  in which the indexes are physically clustered in storage. 
     
     
         46 . The method of  claim 24  in which the concurrent processing is done in a predetermined relative order with respect to the stored data and the replicas. 
     
     
         47 . The method of  claim 24  in which the stored data includes data items of the database that comprise objects in an object database. 
     
     
         48 . The method of  claim 24  in which the stored data includes data items that are provided as objects to an object-oriented application. 
     
     
         49 . The method of  claim 24  in which an object relational broker provides persistent storage of objects for an object-oriented application. 
     
     
         50 . The method of  claim 24  in which maintaining the consistency includes maintaining and making available information about temporarily incomplete states. 
     
     
         51 . The method of  claim 50  in which the incompleteness of the states is communicated to processes that seek to update at least one of the items or replicas. 
     
     
         52 . The method of  claim 24  in which the data items or replicas are accessed from an object-oriented application. 
     
     
         53 . The method of  claim 24  in which two data items or replicas are stored in two different clusters and the consistency is guaranteed by maintaining relationship roles associated respectively with the two data items or replicas. 
     
     
         54 . The method of  claim 24  in which there are more than two data items or replicas and the relationships among them are binary, ternary, or N-ary. 
     
     
         55 . The method of  claim 24  in which there are more than two data items or replicas and the relationships among them are one-to-one or one-to-many or many-to-many. 
     
     
         56 . The method of  claim 55  in which the relationships among the data items and the replicas are bi-directional. 
     
     
         57 . The method of  claim 24  in which relationships among the data items and the replicas are maintained by representing the relationships by roles, each of the roles being associated with one of the data items or replicas. 
     
     
         58 . The method of  claim 24  in which relationships among the data items and the replicas are established by concurrent processing with respect to the data items in the replicas. 
     
     
         59 . The method of  claim 24  in which relationships among the data items in the replicas are created by creating a set of role objects, one for each of the data items for which a relationship is to be created, the role objects including pointers to one another. 
     
     
         60 . The method of  claim 59  in which the creations of the role objects are synchronized. 
     
     
         61 . The method of  claim 59  in which the role objects include replicated information about data items or replicas with which the data item or replica associated with the role object has the relationship. 
     
     
         62 . The method of  claim 61  in which the role object includes a version number that is incremented each time the associated data item or replica is updated. 
     
     
         63 . The method of  claim 61  also including sending messages among the role objects to keep replicated data and version number information current among the role objects. 
     
     
         64 . The method of  claim 61  in which role objects of relationships store version numbers of the other role objects with which they have relationships. 
     
     
         65 . The method of  claim 61  in which the role objects which have relationships store information about missing versions for the other role objects with which they have relationships. 
     
     
         66 . The method of  claim 61  also including assuring that, when a relationship is deleted, the deletion is correct notwithstanding multiple pending delete requests from different objects in the relationship.