dinsdag 9 maart 2010

SQL Zaterdag

Last Saturday i went to ‘SQL Zaterdag’ seminar and this new seminar is great way of being informed about the (new) features of SQL Server. This seminar was organized by SQL Pass.nl, Infosupport and Microsoft Netherlands. There were two tracks: one BI track and a Not BI track:Engine track. I joined the BI track with an old co worker. These were the subjects of the BI track:
  • SQL Server – de BI release,
  • Master Data Management,
  • SSIS : Tuning for Performance 
  •  Building a dashboard with Powerpivot, Reporting Services and Performancepoint .

I would like to share some interesting stuff:

BI release (Hans Geurtsen)
  • The release date of SQL Server 2008 R2 is may 2010.
  • There are really neat features in Reporting services specially for supporting the Selfservice BI strategy of Microsoft. Report builder 3.0 will aid the power user more and more with self supporting functionality.
  • Now you have shared datasets which you can share between reports. So i can imagine that a Power user or developer build a query and when it's authorized and approved it can be shared between reports. So this will lead to less double queries if adopted in a proper way.
  • Shared report parts for mash up support. Mash up technologie will be a technology which will emerge in the coming years. Building in the right way organisations will be more agile because report parts can be created, dropped and changed easily. Another aspect is that when reports use the same reports parts they can be updated with the new report part. But updating the reports in a push strategy is not going to work (at this moment). You can't update all the reports all at once. Yoú'll have to load the reports and refresh the report parts manually.
  • 2 extra pilars in the BI stack for Master Data management (stratature) and for Streaminsight (CEP).
  • Adjustments in the SSAS : Classic OLAP and Powerpivot support.
  • A new edition : Parallel datawarehousing (DatEllegro) .

Master data management (Marcel Westra)

Applications needs to be adjusted for supporting the MDM because data needs to be pushed to the MDM hub and down to the other worker hubs. So it’s gonna take a while before a successful MDM project is implemented. But when this implemented in the right way (?) the source systems will transform into a fact generator. This will have a great benefit in dataquality and will have consequences for implementing and maintaining datawarehouses.

SSIS : Tuning for performance (James Rowland-Jones):
James explained all kind of performance issues with SSIS. Stuff i haven't seen before like the log events screen. I have to look into that. Also he spoke about jumboframes(?). Below you see point that i managed to grab during his fast speed speech:
  • A lot of performance counters which can aid tuning. Too many to write down here. These are familiar and can also be found in the Microsoft certification books.
  • Some discussion about properties which can aid performance like FastParse.
  • Make a baseline when importing data. (use the trash destination).
  • Don’t use ADO.net for your source and destination. It will generate procedural based loops. Used OLEDB or SQL destination.

Building a dashboard with powerpivot, Reporting services and sharepoint (Kasper de Jonge)

Kasper showed great features with reporting, powerpivot and sharepoint. It seems to me that Powerusers, developers, etc will built components with reporting and powerprivot and than they will be uploaded to sharepoint and can be used by the users. Kasper showed all kind of possibilities and I was impressed by the new features. This demo show the integration of powerpivot, excel, reporting services and sharepoint in a great way.


Geen opmerkingen:

Een reactie posten