OWB Implementation Testimonial

1
OWB Implementation Testimonial

• Presented by
• David Cordas
• Data Architecture Team Lead
• LUCRUM, Incorporated
• 312 Plum Street, Suite 1110
• Cincinnati, OH 45202
• 513.241.5949
• www.lucruminc.com

2
Topics

• Client profile
• Architecture layout
• OWB standards
• Job scheduling/Monitoring
• Future state

3
Client Profile

• Cinergy Corp.
• Provider of gas electric serving approximately
  2 million customers in OH/KY/IN
• Initiative began in early 2003 to consolidate
  numerous data silos
• Employed Lucrum to aid in the creation of an
  enterprise wide data warehouse
• Using both Microsoft and Oracle databases

4
Architecture layout

• Hardware 3 Sun Fire 4800
• Dev 2 1.4 GHz processors, 16 GB RAM
• Test/Prod 4 1.4 GHz processors, 16 GB RAM
• Operating system Sun Solaris 8
• Database Oracle 9.2.0.3.0
• Oracle Warehouse Builder
• Dev 9.2.0.2.8/9.2.0.2.0
• Test/Prod 9.0.4.10.0/9.0.4.0.27
• SAN backbone running on Clariion Symmetrix

5
Architecture layout cont.
Dev
Prod
Test
Dev Synch
Design Repository
Design Repository
Design Repository
UAT Promote
Promote
Prod Fix Synch
UAT
Test
Prod Fix
Prod Fix Promote
Deploy
Deploy
Redeploy
Deploy
Deploy
Deploy
Runtime Repository
Runtime Repository
Test Runtime Repository
UAT Runtime Repository
Prod Fix Runtime Repository
Product Release
6
OWB Standards

• OWB configured for 17 concurrent developers
• Common module for conformed dimensions across all
  development paths
• Separate modules for individual project
  development
• Data sources defined by OWB administrator, not
  developers
• No outbound reconciliation!!!!

7
OWB Standards cont.

• Fact table mappings configured for set based
  rather than row based processing
• Dimension table mappings configured for set
  based fail over to row based processing
• Process flows used to organize logical groups of
  work
• Common processing tasks grouped in re-usable
  process flows

8
OWB Standards cont.

• Oracle sources handled via database links
• Non-Oracle sources handled via flat files
• Flat files configured as External Tables
• Custom PL/SQL written outside of OWB and imported
  into meta data
• Objects currently deployed Mappings Process
  Flows

9
Job Scheduling Monitoring

• Cron used to schedule nightly processing
• Data validation checkpoints throughout ETL
  processing
• Parameterized stored procedure used to launch OWB
  process flows
• Process flows used to handle execution of
  dependent sub-processes

10
Job Scheduling Monitoring cont.

• Utilize Runtime Audit Browser (RAB) and/or
  Workflow Monitor for information on OWB
  processing
• OWB currently processing approximately 4 million
  rows of data nightly from 3 source systems

11
Future State

• Implement OWB templates
• Provide robust job notification facility
  (email/paging)
• Implement OWB data validation plug-ins (Fact
  Dimension thresholding)
• Automate OWB meta data feed to custom meta data
  system
• Consolidate 12 additional source systems

12
Questions???

• Thank You!
• David Cordas
• Lucrum, Inc.
• dcordas_at_lucruminc.com