General Enhancements¶

The following general enhancements have also been made to the product:

Improved Handling of Static Datasets¶

Work has been done to improve the efficiency with which static datasets are handled within the system. Many components (such as the Appointment Booking Engine) do not need to process these datasets at all, while others (such as the DSE and Schedule Broadcast Manager) only need to process them for a limited time. The process flow has now been improved to ensure that components only process these datasets where necessary, so reducing the overall load on the system.

In addition, the Schedule Query Manager is now able to 'lazy load' static schedules. Rather than proactively loading the dataset into memory, the query manager will wait until a query is made against the dataset before loading it, and will then hold it in memory only until that memory is required for another purpose. This further reduces the load on the system, especially where static datasets are only likely to be viewed on an occasional basis.

Combined Segmentation Mechanism¶

The Schedule Distribution Service can be used to segment large scheduling problems into smaller internal datasets for multiple DSEs to process. There are three methods of segmentation supported: spatial, time-based and constraint-based. In previous versions the Schedule Distribution Service would evaluate each of these options and pick the best, but in 6.3 it can use these options in combination. For example, a scheduling problem could be segmented both spatially (based on the locations of the resources) and by time (perhaps into separate weekly schedules). The Schedule Distribution Service will choose whichever combination of methods is most appropriate for the given scheduling problem.

See the Architecture and Sizing Guide for details.

Atomic Datasets¶

The Schedule Distribution Service is used for large scheduling problems to filter and segment the input data sent to the DSE and so improve DSE performance. However, for smaller scheduling problems this previously involved a small overhead since some additional copying of data between internal and external datasets was required. In the 6.3 version this overhead has been removed through the use of 'ATOMIC' datasets, which are used by the Schedule Distribution Service to indicate that a DSE can process the original dataset directly. This means that the Schedule Distribution Service can now be used on a multi-dataset system without any negative implications.

See the Architecture and Sizing Guide for details.

Azure URLs¶

After the initialisation of a new cloud deployment has completed, the connections text file that is generated will now also include the URLs for both the new workbench and the Scheduling Gateway. See the Installation Guide for details.

Performance Comparison¶

The scheduling performance of the 6.3 version has been measured against the 6.2 version and shows no measurable difference in performance. In addition, the performance of the appointment booking process has been measured, and the results are included in an appendix to the Architecture and Sizing Guide.