Automatic Partitioning for Improved Scalability

SAP Sybase Event Stream Processor 5.1 SP04 provides the ability to automatically partition elements within an ESP project for parallel processing, allowing greater project scalability by eliminating bottlenecks.

Automatic partitioning is the creation of parallel instances of a given element and partitioning input data into these instances. Partitioning data this way results in higher performance as the workload is split across the parallel instances. Use automatic partitioning to improve the performance of a given element and complex projects, which perform computationally expensive operations such as aggregation and joins.

Quickly and easily create parallel instances of a delta stream, stream, window, or module by using the PARTITION BY clause. You can use either ROUNDROBIN, HASH, or CUSTOM partitioning functions.

You can manually add this clause to your project or use the Visual editor in the ESP Studio to add and configure a partition policy.
  • For more details on automatic partitioning and the PARTITION BY clause, see PARTITION BY Clause in the Programmers Reference guide.
  • For details on how to partition an element using the ESP Studio, see Creating a Partition in the Studio Users Guide.
  • For guidelines and tips on partitioning elements, see the Automatic Partitioning section in the Programmers Guide.
  • For guidelines and examples of how to improve performance of large ESP projects, see Distributing Load Through Parallelization in the Programmers Guide.
  • For examples on using partitioning, see Performance Tuning Using Partitioning, Partitioning a Module on Load, and Declaring a Partitioning Global Parameter in the Examples Guide.