question on analysis services 2005
-
My company is going to build/implement a data warehouse. I'm interested in real world experiences and clarification on the 2005 product. I'm the dba that will run it.
Do you typically need another server for AS or run it on the data warehouse server? The server will be new equipment (not anything old).
Does 2005 have wizards for data cleansing (like names, postal lookups, etc)? Or is this usually code that is written and plugged in?
Does 2005 DTS have wizards for creating cubes, dimensions, materialized views from load data?
Do you typically run Enterprise server so more than 2 gig of ram can be used for the database/AS?
Thanks,
Kevin
-
Kevin
Sounds like a cool job, The AS should run fine on your new server.
This is however dependant on the record numbers.
Cubes are normally built at night so unless you are 24/7 should not be aproblem.
DTS is no longer in 2005 it is SSIS Sql Server Integration Services.
And there are a number of wizards.
Enterpise manager is no longer either. It is the Biz Intell Studio.
Footprints are not a problem any longer.
-
Kevin
Sounds like a cool job, The AS should run fine on your new server.
This is however dependant on the record numbers.
Cubes are normally built at night so unless you are 24/7 should not be aproblem.
DTS is no longer in 2005 it is SSIS Sql Server Integration Services.
And there are a number of wizards.
Enterpise manager is no longer either. It is the Biz Intell Studio.
Footprints are not a problem any longer.
-
Some other neat/cool features to consider:
- SSIS (ie the new dts) can let you load directly to a partition, ie no need to 'process' the cube
- SSAS (ie analysis services) now has proactive caching, so there is no real ned to process your cube again if you don't want to. A simple scenario for this is that you're acceptable latency is 30 minutes, you process the cube for the first time, users hit the cache (ie cube) but the data in the underlying DB changes, once the latency threshold is close/passed, AS will start to pass ueries back to the underlying datasource and rebuild the cache (automagically), when the cache is rebuilt user queries are redirected back to this again. The 'real time' capabilities of AS2005 are quite impressive (on paper)
- Probably better for POC than production efforts, you can build a cube without any datasource actually existing. So build you cube (fact/s and dimensions) and then let AS build the DB schema for you (and IS packages to load this schema I believe)
If only we all could be using AS2K5 on our projects
BTW, if you go up to http://www.microsoft.com/sql , you should find some links to whitepapers that are definitely worht a look (eg processing architecture of AS2K5).
Steve.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply