I attended the SAP Business Objects BI/EIM 4.0 release event in London yesterday.
What really impressed me was the emphasis nearly all of the presenters placed on having a sound data management strategy within an organisation (I’ve been to many vendor events over the years where the technology is sold as the silver bullet for an organisation’s data issues).
A quote used in one of the presentations – on the need for business definitions – particularly resonated with me:
“Lack of consensus on what data means keeps business analytics peripheral to decision making.”
On the technology side two products/areas were of particular interest:
Information steward. First impressions from a 30 minute demo? This looks good and is definitely something I will do further research into.
It seems to have 3 core areas*:
- Data integration, profiling and quality*. Lots of functionality/’dashboarding’ around setting up data rules and profiling data. The data quality KPI dashboard for key business entities eg customer, products etc. was impressive.This particularly resonated with me as providing data quality ‘statistics’ is something I have been working on recently – though not to the level this tool seems to have.
- Metadata management. This includes the abiltity to carry out impact analysis on changing an information asset eg what upstreams systems source the data and linkages to downstream systems/reports that consume the data. All displayed in a nice UI.
- Business definitions glossary. This area was not included in the demo. But, from what I could gather, this can be used as part of overall enterprise definition management process eg making it easier for subject matter experts to keep them up todate – with the ability to link definitions to physical data sources.
* The notes I took have this all as one area but I suspect that it is actually split into 2.
SAP High-performance Analytic Appliance – HANA. Current business information/analytics infrastructure might be typified as:
OLTP – ODS – data mart(s) – data warehouse(s)
Lots of infrastructure, ‘plumbing’ and copying/moving of vast quantitites of data – often required for for performance reasons.
HANA is SAP’s move to exploit the advantages of in-memory computing. Rather than having the data on disk – it is now possible to process massive quantities of data in real time, in-memory, in a much more performant way. This will remove the need to have lots of different environments for performance reasons. One environment for both transactional and all reporting/BI/analytical requirements.
It seems like a generational shift in physical database ‘storage’ technology and would seem to offer benefits in terms of having real ‘real time’ BI/BA and the reductions in infrastructure..
It could also have an impact in other areas of data architecture.
No more excuses such as ‘denormalise for performance’?
With this ‘generational shift’ might there also be a move to using a true relational database management system eg Rel, with finally, a true separation of physical and logical concerns?