Abstract: A method and apparatus for replicating data among sites is provided. The method allows changes to the same body of data to be replicated synchronously to some destination sites and asynchronously to other destination sites. Such mixed propagation configurations allow synchronous updating to selected remote copies of replicated data where data integrity is a high priority, and asynchronous propagation to remaining copies of replicated data in order to allow transactions to be committed locally regardless of whether the transaction is committed at a remote copy of the replicated data. The propagation mode is selectable by a user on a site-to-site basis for each replicated body of data, thus providing a mix of availability and consistency that is not possible in purely synchronous or purely asynchronous replication systems.

Abstract: An apparatus and method for providing adaptable and configurable conflict resolution within a replicated data environment is disclosed.

Abstract: A method and apparatus for replicating data in a computer system having a plurality of sites is provided. Data stored at a first site is organized as one or more sets of formatted data. A mapping is created between subsets of data within the sets of formatted data and one or more object groups. These logical mappings to object groups allow distribution of the subsets of data at a finer level of granularity than previously provided and allow the subsets of data to potentially span multiple sets of formatted data. For each of the object groups, a replica of each subset of data that is mapped to the object group is created at one or more replication sites. When a data modification is detected to a subset of data that has been mapped to one of the one or more object groups, the data modification is propagated to the corresponding replication sites.

Abstract: A distributed system modeler for modeling a distributed system architecture. The distributed system modeler includes logic for receiving interdependency information specifying interdependencies between a plurality of data objects and programs, logic for defining a plurality of modules, wherein each module of the plurality of modules represents a different grouping of a portion of the plurality of data objects and programs, logic for assigning the modules to nodes in a distributed system, and logic for assigning a distributed transfer methodology between modules in the distributed system. The distributed system modeler includes logic for generating impact analysis criteria defining a quality of said distributed system. The impact analysis criteria includes performance metrics, conflict metrics, data object or program availability metrics, and transaction consistency metrics.