I am still trying to get my head around the UDM concept. In particular I am
interested in the following questions I cant see to answer:

1. Is the UDM a big cube?
2. Can I access the UDM without having a cube, i.e for a relational system?
3. Is a perspective a type of cube?!
4. How do all these things relate, especially when looking back to AS 2000?!

Sorry if this things seem vague!!!
--
Simmo

A good reference on the UDM can be found at:
http://msdn.microsoft.com/library/de...IntrotoUDM.asp

In general the UDM is a bridge between the different data sources and the
client tools that access them. Saying that, the MS solution to the UDM is
to utilize Analysis Services and providing various cubes (vs. just one big
cube).

--
HTH!
Denny Lee
<dennyglee_at_hotmail_dot_com>

Blog at:: http://spaces.msn.com/members/denster/



"Simmo" <Simmo (AT) discussions (DOT) microsoft.com> wrote

Thanks Denny I have read this article before. However you cant seem to access
the UDM unless you actually build one or more cubes? Therefore the UDM doesnt
seem to be accessible unless done as an OLAP/cube style structure? Any more
ideas?
--
Simmo


"Denny Lee" wrote:

See
http://spaces.msn.com/members/cwebbbi/Blog/cns!1pi7ETChsJ1un_2s41jm9Iyg!325.entry
....for my view on this matter...
--
Blog at:
http://spaces.msn.com/members/cwebbbi/


"Simmo" wrote:

Thanks Chris that blog makes a lot of sense. Do you or anyone else have any
thoughts on perspectives and DSVs and what the hierarchy of it all is, I cant
seem to get a definitive picture...

Cheers
--
Simmo


"Chris Webb" wrote:

A perspective is like a view on a cube - it allows you to make certain
measures, attributes, dimensions etc invisible to reduce the complexity of
the cube and make it more usable - some dimensions will have hundreds of
attributes, for example, which won't all be interesting to all users and
which would otherwise clutter up the UI of whichever client tool your users
are using to browse the data. Perspectives are not a form of security though.
They appear as cubes to users and behave in the same way.

A data source view is a way of filtering down the number of tables and views
in a relational data source to make the process of building cubes and
dimensions easier. In many cases there might be hundreds of tables in a data
source and this makes selecting the relevant ones when you're going through
the cube/dimension creation wizards pretty tricky; in the dsv you just need
to select the relevant tables once, and then you have a much shorter list to
choose from later on. Dsvs also allow you to do stuff similar to what you're
able to do with a view in a relational database (eg create calculated
columns, named queries) for when you're trying to build a cube but don't have
the permissions you need on the relational source data.

HTH,

Chris
--
Blog at:
http://spaces.msn.com/members/cwebbbi/


"Simmo" wrote: