Posted by Doug OKeefe on Fri, Nov 13, 2009 @ 12:54 PM
Written by: Kevin Barth
Whether it's new music or new software, you can rely on consumers looking down the road to the next big release while the current one is still warm in their hands. And so it is with Essbase 11.1.1.3. While Oracle hasn't yet committed to a release date, they did whet our appetites with some enhancements for 11.1.2 at OpenWorld in October.
It's pretty clear that the Aggregate Storage Option (ASO) figures heavily in 11.1.2. ASO will feature functionality like that of the procedural calc engine used now in the Block Storage Option (BSO). What does that mean for geeks like us? It means you can run a calc script in ASO (without MDX) and the data can be stored directly in the cube. There may be some limited functionality within the calc engine initially, but it's a heck of a start. ASO will also support the predefined allocation function. Lastly, the calc engine will be exposed within the Java API and MaxL.
From an enterprise-ready point of view, there are a few enhancements that, while they're not terribly sexy, are definite improvements from an infrastructure standpoint. These include process monitoring, status dashboards, identification of key metrics, centralized logging, and error identification. Oracle also revealed some Essbase Studio improvements, such as the ability to move, copy, and backup the import/export catalog. Additionally, there is increased support for ODBC DSN sources and improved data source synchronization.
If you're looking down the road already, these enhancements show all the promise of improved performance and usability.
Questions or comments?
info@crownpartners.com
Posted by Abe Combs on Tue, Oct 20, 2009 @ 11:55 AM
-Written by Abe Combs
FDM will work with aggregate storage (ASO). It also works with Internet Explorer, Adobe Acrobat, Notepad, and the newest KitchenAid food processor.
Okay, sarcasm aside, FDM has a little-used feature call mShellandWait-a subroutine that lets you call pretty much any external executable code, whether it's a batch file or an executable. Even a batch file you would use to call a MaxL process that builds or loads an aggregate storage cube.
But, the operative words here are "will work", which are different from "is supported". While FDM does not support aggregate storage, you could use the mapping and import formatting (both strengths of FDM) to get a file all the way through to export where you would then launch a MaxL process (via a batch file) that loads and aggregates an ASO cube.
No, it's not ideal; you're flying blind to a degree. You have no interface into the executable you launch, so you can't be sure it was successful. But, if calling MaxL is your goal, and you're a little creative, you can use MaxL's ability to write error logs and use FDM scripting to read the logs and send emails-and in the process make up for a lot of the shortfall.
If you're still with us, here's even more excitement: you can use FDM to call any executable-so in some ways you're limited only by your imagination. One of our favorite uses of the mShellandWait command is in conjunction with Hyperion Reports. Hyperion Reports' batch scheduler actually allows you to execute a report from a batch file on the fly, and email the results to a select audience. So once FDM has sent you the email confirming successful execution of your cube/HFM build process, you can send yourself an email, using mShellandWait and Hyperion Reports, that checks the database for validity. Did the balance sheet really balance? Do those alternate roll-ups tie out? Or, you can be a real hero and send the accounting department the first-look P&Ls during close week.
Questions or comments?
info@crownpartners.com