Here is a useful page from the SQL Server 2008 BOL that summarizes the SSAS performance improvements in Katmai pretty well. It also shows which functions are optimized. For example, the Lag function wasn’t optimized in 2005 but now is.
Archive for month: November, 2007
In my opinion, the biggest challenge the Microsoft BI initiative faces today is the inadequate support for Analysis Services. The premimum Microsoft reporting tools claim to support SSAS but they all take an idependent view about what features they should support or not. To make things worse, even supported features turn out to be not “supported” enough and subsequenly declared as “by design” or known limitations. This cripples UDM and forces customers to look outside the Microsoft BI stack, hoping to find third-party tools to fill in the gap.
Recently, I had the pleasure to take a look at the pre-release bits of the Radius Producer by 90 Degree Software. Those of you who installed the November CTP of SQL Server 2008 and played with the standalone Report Designer, will undoubtedly find many similarities between both tools, in terms of artchitecture, designer experience, even look and feel. One feature that Radius Producer excells in is end-user reporting from SSAS. For example, the snapshot shows the Radius Producer filtering support which I particularly liked. The user can filter on multiple dimension and measures and the tool provides common filter conditions that business users would appreciate, such as top, bottom, greater than, less than, etc. When the user selects the Ask Me Later option, the filter is promoted to a report parameter. Based on what I’ve seen, Radius Producer generates very clean and efficient MDX although I am yet to test with larger cubes. I was able to produce reports succesfully from both SSAS 2005 and 2008.
Radius Producer is more ambitious than being a reporting tool only. It promotes colleboration among information workers. Users can save queries and report snippets, and subsequently share them on the Radius network. Think Zune but instead of music you share report gadgets. If you are on a lookout for a third-party reporting tool that generates RDL and provides a good support for SSAS, take a moment to evaluate the forthcoming release of Radius Producer.
The November CTP (CTP5) of SQL Server 2008 is out. Here are the most significant BI-related highlights that caught my attention.
The standalone Report Designer has undergone a major facelift. It now sponsors a shiny Office 2007 ribbon interface. It will surely charm end users who are planning to use the standalone Report Designer for ad hoc reporting. The bad news is that there is still much integration work ahead. For example, the Analysis Services and custom data processing extensions are not integrated yet. Further, only the Dundas chart component has made the CTP5 cut. Nevertheless, I am very excited about the standalone Report Designer and its long-term potential to unite professional and end-user reporting needs.
The block computation enhancements debut in CTP5. Based on my experiments, they result in 50-60 percent and more improvement in query response times. For example, a highly inefficient Report Builder query that filters on a measure would take hours with SSSAS 2005. With CTP5 of SSAS 2008, the same query finishes under a minute! Put in highly technical terms, the nastier the query is, the faster it will with SSAS 2008. There are also optimizations in MOLAP write-back that remove the need to query ROLAP partitions although I haven’t given them a try.
CTP5 brings lookup performance enhancements. Are the SSIS guys lazy or what? J
One BI-related enhancement with a great potential is Change Data Capture (CDC). When enabled, this features tracks changes to data, including inserts, deletes, and updates. This means that ETL process can quickly detect what changes have been made to a given table by just querying the CDC internal tables. For example, you query CDC to return only the new records that have been added to a table after a given data. As you can imagine, this baby will be a welcome enhancement with large data warehouses and can help you optimize the ETL processes.