Dependency tracking application - sqlserver-olap
This is a discussion on Dependency tracking application - sqlserver-olap ; Hello, I need advice on whether SSAS or OLAP in general is a good choice to model dependency chains. I need to model and store application depdendencies e.g app A is dependent on database db1 which resides on server 1 ...
![]() |
| | LinkBack | Thread Tools | Display Modes |
|
#1
| |||
| |||
| I need advice on whether SSAS or OLAP in general is a good choice to model dependency chains. I need to model and store application depdendencies e.g app A is dependent on database db1 which resides on server 1 Db1 is also dependent on ETL job Load1, which in turn is dependent on a data file df1. The goal is to be able to build a report so user picks an object (file, db, app, or server) and can view it's dependency chain. e.g user chooses to view dependencies of data file df1 then the report should list all it's dependent objects- File df1 affects, Load1, which in turn affects Db1, which impacts App A. Thank you for your input. Ken |
|
#2
| |||
| |||
|
On Nov 14, 4:15*am, Ken > Hello, > I need advice on whether SSAS or OLAP in general is a good choice to > model dependency chains. I need to model and store application > depdendencies > e.g app A is dependent on database db1 which resides on server 1 > Db1 is also dependent on ETL job Load1, which in turn is dependent on > a data file df1. > > The goal is to be able to build a report so user picks an object > (file, db, app, or server) and can view it's dependency chain. e.g > user chooses to view dependencies of data file df1 then the report > should list all it's dependent objects- > File df1 affects, Load1, which in turn affects Db1, which impacts App > A. > > Thank you for your input. > > Ken Ken, this is the BI metadata handling we are all waiting for years! Using existing technologies, we suggested a way to track this dependencies adapting the development methodology to the existing tools. Let's take a look at our SQLBI Methodology (http:// www.sqlbi.com/sqlbimethodology.aspx). Marco Russo http://www.sqlbi.com http://sqlblog.com/blogs/marco_russo |
|
#3
| |||
| |||
|
Marco, please point me to where on this site, you recommend (www.sqlbi.com/sqlbimethodology.aspx), can I find information relevant to my query? Thank you |
|
#4
| |||
| |||
|
Ken, Maybe I'm not getting what you're asking in total, but this seems one:many, but maybe with distinct hierarchies; you may be able to do this with a regular dimension (with empties) or parent-child dimension. This is an pretty original use of OLAP though, I like it. We did something like this with a troubleshooting app that built paths which we did by building parent-child dimensions that mimicked the decision trees. Yes, form what I understand, OLAP would handle this, but I'm not sure what the Facts would be, maybe the files them selves with the files sizes or other details being the measures. With that maybe I can get you to think of it in a differnet way, though. One of the paths seems like a AppUsageDependency, that is what it is required to actively Use the app. The other path you enumerate is a AppPrepDependency, that is, what is required to ready the application prior to using it (loading the data , etc.) If you have one of the hierarchies like so, likley called AppUsageDependency, I might turn it upside down. An app cannot exist sans a server (or client). So would the server be the leaf or root? Baring that, I think the server would be at the root, and the apps would be the leaves. You may even want to include the individual files (DLLs, other components, etc) that the app requires, but not sure if you need it that detail. (I might even suggest server instance as well as an additional level in there) AppUsageDependency: Server -> DBServer -> DBServerInstance -> App I might also suggest having attributes at the server level in there like for the Virtual or Physical (booleans), and IPAddress. AppPrepDependency: Server -> DBServer -> DBServerInstance -> ETLJob -> SourceFile Usage: When they browse the dimension (in the report no matter what UI you end up using), they could start at the bottom and work their way north until they get the root, or the other way around, start on top and work their way down. Or you could use in your MDX query the Descendants and Ancestors function would show you the complete path from root to leaf given any root or leaf, or for the each of leaves. -exologic |
![]() |
« Previous Thread
|
Next Thread »
| Thread Tools | |
| Display Modes | |
| |
All times are GMT -4. The time now is 11:46 PM.




Linear Mode