Live Blogging Data Integration in Large Organizations

May 30, 2010

This is live blogging from the Donald Farmer seminar from the SQL & BI convention in Tel Aviv. I would really like to thank Hamada who lent me his netstick so I can blog – Thanks Hamada!


Business Intelligence is about people not about IT.
Your data is always connected to real world and in direct connection to it. It’s important to know your business for you to design a good Decision Support system.
BI has always been in the middle of the spectrum – tactical decisions. CEO and operational workers will usually look just at your BI reports, they won’t create them themselves.


For the CEO, you have strategic decisions – Dashboards and Scorecards. Getting data aggregated.
For Operational Workers, you need embedded reports, analytic APIs so that they can get the info from where they’re working (no special knowledge needed). Getting the data to the right place.


SSIS for the operational level. SSAS to model the operational data giving a consistent view of the data for the entire organization. PerformancePoint Services giving the high level strategic view.


Decisions flow from the strategic level down to the operational end. Traditional BI from the tactic level may move up to the strategic level and change their decision.


Questions you need to ask yourself: Who am I giving the info to? What level of info do they need? How will they use the info to make decisions?


Prototype a BI project. Test the accuracy and performance of the DW and also see that people actually use the solution you developed. Do Beta testing to see how people use the reports – look at the query log their using and tune the DW according to that.
An outer user should be treated differently only from the security aspect, but except for that you treat them just as any other user. This is unless you just give them a report rather than let them use freely the data.


Operational – won’t learn new application – reports, embedded functions, mobile BI
Tactical – reports, cubes, annonationas and SharePoint workflow.
Starategic – PerformancePoint Services, annonationas and SharePoint workflow.


SharePoint is the largest growing server in Microsoft. SharePoint is growing because people need to share info and work on it together in a managed way. SharePoint 2010 has social features as well, as collaborating through social networks has become more and more important. Where you’re collaborating and where you’re sharing, is now also where you use your Business Intelligence, which is also integrated into SharePoint.


1. IT oversight – how much control IT has


2. Organizational alignments – coordination with the goals of the organization


3 S of a DWH:
Scale
Standards – built to. Critical. Laws for instance about HR DWH (for instance) and who do you allow against it (Security).
Skills


DWH and Reports – high on 1 + 2
Spreadsheets & desktop databases – low on 1 + 2


Between the two there exist:
Shared sheets, Databases
Departmental Data Marts


Sharepoint as a way to bind all of these together.


New to SQL Server 2008 R2 – Master Data. Master Data is not a DW. DW is for analysis and hold data of current state + history. Master Data is about the current state of the Business Entity – is for operational use. Master Data – keeps the data synchronized for all the departments.


Next version of SQL Server in 2011 will include more Data Quality capabilities based on abilities acquired from Zoomix.


ETL – Extratct Transform Load. Moving data in the organization. SSIS.
EAI – Enterprise Application Integration. Moving messages in the organization. Biztalk.


Biztalk and SSIS work both on 3 layers: Data, Logic, Presentation. Coordinate on all levels.


Request and Reply – with ETL updates. Cross reference table is updated by an ETL process in batch mode.

Add comment
facebook linkedin twitter email

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

*