Hi All,
New to the world of COGNOS BI and apologise if this question has already been asked. Is it possible to take advantage of stargateviews created in TM1 to be utilsed in COGNOS BI? Within COGNOS BI is it possible to link in a report to the stargate view and then take advantage of this pre-cached performance??
If so, can someone please kindly provide a documentation link?
Cheers
Zark
Stargate Views and Cognos BI - Performance
-
declanr
- MVP
- Posts: 1831
- Joined: Mon Dec 05, 2011 11:51 am
- OLAP Product: Cognos TM1
- Version: PA2.0 and most of the old ones
- Excel Version: All of em
- Location: Manchester, United Kingdom
- Contact:
Re: Stargate Views and Cognos BI - Performance
It does it automatically. If the view opened from a BI report is within the constraints of TM1's VMM/VMT parameters it will build a stargate view and if when opening a report in BI that looks at TM1 data there is already a stargate view (created due to a previous BI connection or previous TM1 connection) then it will utilise that. The TM1 server doesn't really differentiate how the data is being accessed; be it from BI, perspectives, TM1 Web, Contributor, Insight etc.
The differences in speed between viewing medium is all down to distance to the TM1 server and also how the end tool (e.g. BI) utilises that data.
You should also note that you can utilise BI caching which can be turned on or off in the FM package.
The differences in speed between viewing medium is all down to distance to the TM1 server and also how the end tool (e.g. BI) utilises that data.
You should also note that you can utilise BI caching which can be turned on or off in the FM package.
Declan Rodger
-
7Zark7
- Posts: 81
- Joined: Tue May 31, 2011 6:29 am
- OLAP Product: TM1 and EV
- Version: 9.1 SP3
- Excel Version: Excel 2010
Re: Stargate Views and Cognos BI - Performance
Thanks Declanr
Can the following be confirmed, my understanding of Utilising StargateViews is still a little fuzzy:
1) If I build a StarGate view via function: ViewConstruct ( with appropriate VMM and VMT parameters set and other cfg setting as well defined for that cube ), lets say "DaveTest" targeting year = "2015"
for all cells in the cube will I need to refer to that view in my BI report? Or the fact that the those cells for "DaveTest" have already been cached via ViewConstruct, a user
would be able to take advantage of the caching when opening up any view in the cube via excel / tm1web / COGNOS BI as long as view is referencing year = 2015??
2) If this view is deleted will the cache be destroyed as well? My understanding is now since caching is only effected on server restart and if there are data changes associated with the cube I.E.
TI processes being run attribute changes via a rule hanging of this cube. Object view deletion has no effect on cache?
3) Running a call to a viewconstruct after a cube load would be optimal to re-cache the cube for particular slices of the cube defined by the business. Again, if a user views a set of data from that cube that is
within the cell footprint of the caches views then that data should already be pre-cached.
Cheers
Can the following be confirmed, my understanding of Utilising StargateViews is still a little fuzzy:
1) If I build a StarGate view via function: ViewConstruct ( with appropriate VMM and VMT parameters set and other cfg setting as well defined for that cube ), lets say "DaveTest" targeting year = "2015"
for all cells in the cube will I need to refer to that view in my BI report? Or the fact that the those cells for "DaveTest" have already been cached via ViewConstruct, a user
would be able to take advantage of the caching when opening up any view in the cube via excel / tm1web / COGNOS BI as long as view is referencing year = 2015??
2) If this view is deleted will the cache be destroyed as well? My understanding is now since caching is only effected on server restart and if there are data changes associated with the cube I.E.
TI processes being run attribute changes via a rule hanging of this cube. Object view deletion has no effect on cache?
3) Running a call to a viewconstruct after a cube load would be optimal to re-cache the cube for particular slices of the cube defined by the business. Again, if a user views a set of data from that cube that is
within the cell footprint of the caches views then that data should already be pre-cached.
Cheers