GenODS.. Details

[maheshbabu]

Hi All,
        Can somebody tell me about the functionalities of ODS.. and for what other services it is related to etc., I have searched for it and I got about OCS. Kindly help me in ODS.

[cavagnaro]

ODS retains all stat data which is feed by StatServer, then ETL extracts the info from it to build the Datamart tables.

[Mine]

Each  application's connection is

"Stat Server"→"Data Soucer"→"ODS DB Server"→"DB(ODS)"→"ETLRuntime"→"DB(DataMart)".

you can use one table of DB(DataMart), from some Stat Server put data.

[maheshbabu]

what are all the applications related to it. IS DMA / SCI connected to it... DMA gets refreshed/updated for every 15 minutes. Hope this is done by ODS...

Can u plz explain in detail.. / can I have any docs... !! [attach=#]

[Mine]

where do you watch this error messages ?

please show me the error log.

[victor]

Hi, Mahesh,

sorry, but we cannot distribute copyrighted Genesys manual on this forum.
Tony wrote an excellent manual for what you needthough here:

http://www.sggu.com/smf/index.php/topic,2532.msg9742.html#msg9742

I hope it helps.

Vic

[maheshbabu]

Hi .. I got all those stuff... But I can't find the ODS functionalities & interaction there... Can u help it out..

[tony]

Hmmm... OK;

ODS is the Operational Data Storage which contains Object and Statistical raw data values, prior to them being collated by an ETL process, into a Datamart, where that data is aggregated and used for reporting purposes.

There may be more than one ODS, per Historical Reporting Solution (Call Centre Analyzer (CCA)) - this is dependant on the size/scale of your call centre environment.  The ODS itself is maintaned through Java processeing and User-defined parameters, through "PROPERTIES" Files which are usually co-resident on the Server which houses the ODS.

The Data Migration Assistant (DMA) Application is a Desktop WINTEL Application, where a User determines which Objects, Statistics and Filters should be collected, for reporting purposes.  This information is propogated to the Datamart.  Once a new requirement ("Layout") is applied within DMA, an ETL process creates a new series of ("OL") Data Tables within the ODS.  These Tables will contain the raw data, extracted from the routing via a StatServer, based on the Objects and Statistics defined in the "Layout".  Once the process of setting up a new "Layout" is complete, ODS will begin to collect the Object and Statistical data, as defined in the DMA.

The ODS is designed to extract, transform and load (ETL) data to the Datamart, at a rate determined withn Java Settings ("PROPERTIES" Files) and various other settings which can be applied through the Configuration Manager GUI (CME).  it is normal that data is ETL'd every 15 minutes and, once the data transfer from the ODS to the Datamart has been confirmed, the data is deleted/removed from the ODS (purged).  This process can be monitored in the DMA Applcation.

So, the Datamart is then fed near-real-time data (usually in 15-minute chunks) and it then collates those chunks into Aggregated values (NOAGG, HOURLY, DAILY, WEEKLY, MONTHLY, QUARTERLY, YEARLY). The retention rules for the storage of this data is maintained within the ETL Assistant and it is normal that lower aggregated data (NOAGG and HOURLY) are purged quite quickly (within 7 days), whereas higher aggregated data (DAILY, etc.) are stored for longer periods (weeks, months or years).  Getting the purging wrong can mean that you are keeping 15-minute items of data for a lot of Objects ("Agents" or "Queues") for long periods of time and your Datamart will fill up very, very quickly...

That's about it, in a nutshell... If you have any questions, you should really book yourself on a course...

Tony

[victor]

Hey,

I think Tony pretty much nipped in a bud.

Just think about it this way:

you need to see how call center performed in 15 minute intervals, 30 minute intervals, per-hour, per-day, per-week, right? Statserver by itself has only realtime statistics that use some soft of time reference, such as since last minute, since last 10 minutes, since last day. You cannot ask statserver for statistics from yesterday or last hour or two weeks ago. All it knows is the statistics as it is right now.

So, you need snapshots of those statistics as a day progresses. Snapshot of all statistics at 10:00, snapshot of all statistics at 10:15, then 10:30, then 10:45 and so on. And then you can put those snapshots together and reconstruct the whole picture, right?

You need to store those snapshots somewhere, and ODS is the temporary database where the snapshots are stored in fifteen minute intervals. DataSourcer is the actual program that connects to StatServer and takes those statistics. Think of Datasourcer as a guy with paper and pen who is looking at CCP and writes down all the numbers it sees on the piece of paper. Obviously, ODS is your piece of paper then

Then, you have ETL that takes those 15-minute fragments and pieces them together to make statistics for hourly, daily, weekly and monthly reports and puts it in datamart. Why? Because this way you do not need to do the calculations yourself when you want to see the report - ETL prepares all the possible variation of reports for you in advance.

How does system know what sort of reports you need?
This is where DMA comes in - this is where you tell it what sort of reports you need by defining statistics and layouts in it. This way, system knows what statistics it needs to take a snapshot of, and how to group those statistics together.

So:
DMA: define what sort of statistics you want to be shown in your final reports
DataSourcer: get those statistics from StatServer, or the guy who is looking at CCP and writes down the numbers.
ODS: that piece of paper where the guy is writing down the numbers he sees in CCP.
ETL: your handy macro in Excel sheet that adds all the numbers you wrote down on piece of paper
Datamart: the result column in EXCEL that contains the final numbers you look at

Easy?

[mahesh]

Hi.. Thanks a lot... This is really a gr8 notes...