Introducing Parallel Job Structural Differences

Server Jobs and Parallel Jobs are fundamentally different in the types of Job designs which they can support. In particular, Server stages support a number of capabilities distinct form their Parallel equivalents .

For example, some Server stages …

• support the concurrent reading from and writing to a single data store,

• support multiple input and/or output links, sometimes using different data stores for each link, or

• have no Parallel equivalent

These differences result in Parallel jobs looking structurally different to their Server ancestors. In particular a Parallel job may differ from its Server ancestor by…

• representing a Server stage with a different Parallel stage,

• representing a single Server stage with multiple Parallel stages, or

• representing a single Server job with multiple Parallel jobs, so the inputs to particular stage are handled by one job, and the outputs by another job. These instances also require the generation of a Job Sequence to manage the execution order of these separate jobs.

Finally, note that these structural changes can be applied iteratively, so S2PX may construct a Shared Container (for example) which may itself, due to the nature of its constituent parts, require further decomposition.

All these structural differences are described in the pages in this section.