Abstract: Updating a data storage unit using tenant specific update policies is disclosed. In an embodiment, a plurality of application events from an events publisher is received at a data pipeline manager. The plurality of application events are associated with data at a source data storage unit and are stored at a data pipeline data storage unit. An update process is initiated, based on an update policy associated with a particular tenant. During the update process, one or more application events associated with the particular tenant are selected from among the plurality of application events stored at the data pipeline data storage unit. Data associated with the one or more selected application events is selected from the source data storage unit. Data selected from the source data storage unit is stored at a target data storage unit.

Abstract: In an embodiment, a method is described for partitioning a data table into one or more partitions based on tenant specific partition strategies. In an embodiment, one or more tenant identifiers within a data store are identified. Each of the tenant identifiers identifies a unique tenant. The table of data is partitioned into partitions, and each partition is associated with a particular tenant identifier and comprises data associated with the particular tenant identifier. A particular partition from the partitions is partitioned into one or more sub partitions. A custom sub partition key is determined and associated, at least in part, with particular tenant identifier. In response to a data access query associated with the particular tenant identifier, a data storage unit is queried using the custom sub partition key stored in the metadata table.

Abstract: A data processing method comprising the computer-implemented steps of using a process management computer, receiving an electronic workflow document in response to execution of editing instructions at a user terminal that is coupled by network to the process management computer, wherein the electronic workflow document defines a workflow using a plurality of tags and statements that specify steps in the workflow, wherein each of the steps is either an atomic type step or a non-atomic type step, wherein at least a first step that is atomic comprises a plurality of instructions and is programmed to signal, to a second and successive non-atomic step, normal completion of execution of the first step; wherein at least a second step that is non-atomic is programmed to call a completion callback in response to the signal; using the process management computer, parsing the electronic workflow document to form an in-memory representation of the workflow in computer memory; using the process management computer and usi