For Jedox customers with a strong interest in scalability and performance, 2019.2 is one of our most exciting releases to date. It sets the benchmark for performance of large / complex models, as well as laying a strong foundation for our reporting evolution, Dynamic Views. Here are some highlights of the release.
In-Memory Database: Enhanced Performance Everywhere
We have improved performance on multiple fronts. With Autosave, Jedox now takes care of continuously committing in-memory values to the database filesystem. The save interval (set at a default 5 minutes) can be configured, is completely asynchronous (i.e. it does not stop or block Integrator process, user processes or security). The implication of this constant commit is that the journal state will always be around 5 minutes old. Combined with the ability to turn on dimension and cube binaries only (available since 2019.1), dramatically reduces start times and scheduled auto backup save times on large models.
Faster Login for SSO/ Azure Active Directory
On the server side, the ability to disable continuous fetching of SSO/AD groups means logging in for these scenarios are faster.
Optimized Rules and Selection on Large Dimensions
For users, rule optimization for simple formulas saves memory and speeds up calculations at run time. Performance optimization of web element selectors on large dimensions (read: millions of elements) fills a longstanding request by Jedox users. You can now display huge dimensions easily in element selectors, rule element pickers, Subset Editor and Paste Element dialogs. Note, that this is for web dialogs only. As we are gradually moving all excel dialogs to web dialogs, this will soon be available in Excel too.
New GPU Advisor
The GPU Accelerator Advisor now has a new home in Modeler and the extended functionality allows you to test your cube for suitability for GPU Acceleration. It gives a detailed breakdown of calculation and rules compatibility that can be exported to a file.
In-App Performance Monitoring (in preview)
Our new Request Profiler (in preview) gives an enterprise-level, forensic analysis of server requests per user which can allow for understanding performance bottlenecks around data and rule calculations. Once enabled for a specific user, it can show the slowest requests and analyze the hierarchical dependencies for performance analysis.
Jedox Integrator: From Best Friend to BFF
We’ve made a bunch of performance improvements in one of the most powerful components of the Jedox platform too: Jedox Integrator, . You can now use a new load type CubeSliceCopy to directly send a splashing command such as ‘copy [xx]’ or ‘100 like [xx] ‘. This has large performance implications, especially when moving data around in a cube (for example, copying actuals to forecast version).
When it comes to productizing a Jedox Integrator integration process, cleaning up loose ends can be tiresome. Not anymore. The Remove Unused Components feature displays all unconnected and orphaned components and allows you to select/deselect all or individually and delete in one step.
Finally, the new Show value feature in Variables allows you to see the current value contained within the variable, which is particularly helpful if the value is coming from Groovy or system settings and the current value is not apparent.
Dynamic Views: Exposing Views and Subsets
Giving easier access to cube views and dimension subsets is the first step on our journey to making Views a more powerful tool in reporting. Stored Subsets and Views have been available for a long time in Jedox; now we are making them easier to access. In Jedox Excel Add-in and Report Designer in Jedox Web, stored views can be accessed via a drop list on the toolbar. This allows users to review, edit and save Views quickly. Subsets can be accessed in Modeler and in the Element Selector, again making the change of perspective faster. In the Paste View dialog, “Show now” allows you to preview your selection before you paste into the spreadsheet, much like a pivot table.
Spacers in Picklists for Clearer Sorting
You now can have spacers within your subsets. In 2019.2 you have the ability to not only add spacers but sort the picklist in any custom order you want.
Spacers can be used in subsets, views and dynaranges, but only within the Picklist context.
Lastly, simple changes can sometimes have big impacts. Where possible, we are moving to single clicks for user interactions. You might have already noticed it in dynaranges in 2019.1. We now have single clicks for views (expand and contract), plus element selectors.
So, there you have it. A big release that has big impacts on performance and views, both from a server and user perspective. To review the entire release notes document, visit What’s New in Jedox 2019.2 in the Jedox Knowledge Base.