Legacy Migration using a Virtual : A Case Study Gary Meyers: Anadarko Petroleum Mike Skeffington: EnergyIQ info@energyiq.info www.energyiq.info
Overview Migrating a legacy, distributed data management solution can be disruptive The issues are multiplied by the number of integrated systems To mitigate risk, a migration should be managed in phases that minimize disrupting existing operations and deliver added value to the business This paper will provide a case study of a project designed to: Migrate a complex data management environment over several phases Deliver key business value with each phase Minimize disruption and reduce IT risk with each phase Replace proprietary legacy systems with a COTS solution info@energyiq.info www.energyiq.info 2
10+ Years Of Evolution Well Header Gulf of Mexico North America Production Formation Reference Dir. Surveys Paleo Test Core 3.7 DB Direct Scripts International Completion Injection Scripts Etc. Vendor/Proprietary/Public 10+ Data Sources Primary Data Types and info@energyiq.info www.energyiq.info Corp. Repository 6.1M Well Header Records Dozens of
Design The Future Gap Analysis: Interviewed over 90 business users to understand their data needs Findings: Single trusted repository was needed Expose data blending processes and quality results for risk assessment Make data easily available to business functions to improve decision making Ensure long-term sustainability by adopting a commercial solution Build independence from data providers Challenge: Timely delivery without impacting existing workflows and processes Approach: Iterative phases to deliver critical business value, minimal disruption info@energyiq.info www.energyiq.info 4
The Desired Future State Sources Consumers Sources Consumers Scripts 3.7 Scripts Future State Loader 3.8 Data Access Layer/API Integrated ly Supported Well Hierarchy Data Management Apps (not TOAD) E&P Aware and APIs E&P Data Quality Standard Reports Query / Export Workflow Automation Platform Broad User Application Support info@energyiq.info www.energyiq.info 5
Approach Phase 1 Challenge: has multiple records for the same wells - Very confusing to users Resolution: EnergyIQ software implemented to expose the blended Well Hierarchy Result: Visible data quality, richer data, happy users info@energyiq.info www.energyiq.info 6
Phase I (++) Well Header Gulf of Mexico North America Production Formation Reference Dir. Surveys Paleo Test Core 3.7 DB Direct Scripts? International Completion Injection Scripts Etc. Blended Well Header info@energyiq.info www.energyiq.info
Approach Phase 2 Implement a virtual 3.8 data model as a series of views against the Well Hierarchy and data stores Deploy a 3.8 Gold compliant browser to meet immediate business needs (query/browse data) No changes to proprietary data loaders and existing workflows Browser 3.8 Tables 3.8 Views Result: Mitigate risk for the full migration Immediate value by exposing the views to Spotfire Sets up future migration with minimal impact Proprietary Legacy Well Hierarchy info@energyiq.info www.energyiq.info 8
Key Challenge: Performance Desired Outcome: Browser application performs as if it were accessing a physical database Approach: Populate key reference tables within 3.8 Establish several key views as materialized views Add a new table to provide an index between the two physical data stores, referenced by the views Iterate and tune the view designs in UAT Result: Performant, and will improve with future phases info@energyiq.info www.energyiq.info 9
Approach Next Planned Phase Deploy commercial Well Hierarchy in 3.8 Deploy data applications and management tools Expose data access layer for long term sustainability New Data Access Layer/API Browser Benefits: Reduce operational costs and risks Improve performance Expose common access protocol Replace custom code with commercial Begin migrating consumers to new access layer Begin migrating sources to new 3.8 Proprietary Legacy 3.8 Tables 3.8 Views Well Hierarchy info@energyiq.info www.energyiq.info 10 10
Approach Possible Next Steps Migrate sources to 3.8 with commercial loaders Create a virtual through a set of views preserving legacy applications before migration Continue migrating legacy consumers to the Data Access Layer Benefits: Further reduce risk and cost Retire legacy systems (, consumers, etc.) Retire legacy loaders//scripts Allow the business to migrate apps on their schedule Approach removes data provider dependencies New Data Access Layer/API Legacy Proprietary Legacy Browser 3.8 Tables Views Well Hierarchy info@energyiq.info www.energyiq.info 11 11
Long Term Objective Well Header Sources Consumers Gulf of Mexico North America Production Formation Reference Dir. Surveys Paleo Loader 3.8 Data Access Layer/API Integrated International Test Core Completion Injection Etc. Views Legacy info@energyiq.info www.energyiq.info
Summary And Lessons Learned Large migration projects are often disruptive to the business due to the inherent complexity and timeframes involved Use of a virtual database approach can minimize disruption This approach empowers IT and Business to migrate/upgrade on own schedule Deliver incremental value with short phases Performance is a key concern consider view designs This approach will work with non- sources into a platform Utilization of a Data Access Layer (API) for loading and consuming data should mitigate the impact of future migrations info@energyiq.info www.energyiq.info 13
Questions? Gary.Meyers@Anadarko.com Mike.Skeffington@EnergyIQ.info info@energyiq.info www.energyiq.info 14