Large-Scale Customer Analytics Data Migration
Case Study: Data & Analytics
SITUATION & BUSINESS CHALLENGE
A multinational beverage and foods organization needed to decommission and migrate key business processes and data analysis around customer analytics from its legacy data platform, Teradata, to Oracle’s Exadata platform. The migration was part of a strategic mandate to consolidate on a single data platform and leverage Oracle’s in-memory platform, which closely aligned to the organization’s business needs, planned operational efficiencies and single version of truth for customer analytics.
One of the key challenges was that Teradata had been in use at the organization for many years with business processes and data analysis siloed across the organization in numerous forms, ranging from scattered data repositories to memorization by employees, with no central documentation. The migration would require a thorough understanding and documentation of the entire Teradata functionality ecosystem — business processes, data structures, interfaces and analytics — as well as how to implement this functionality in Oracle’s Exadata. Additionally, more than 13,000 Teradata tables needed migration to the new platform, representing a massive workload for the organization.
Skillfully executing an analytics data migration project of this scale required deep understanding of data and analytics and the methodologies and pitfalls of such an undertaking. The organization searched for a consultancy with the experience and know-how to build and deliver on a viable strategy that would see the organization through the migration. The organization found this expertise in AIM Consulting.
An AIM team composed of a data and analytics subject matter expert, data engineer, and data analyst thoroughly analyzed and documented the organization’s customer analytics requirements, design, and business processes needed to successfully lead the migration. Based on the requirements, new data structures and analytical models for the Oracle environment were developed, including the migration of all data ingestion workflows from Teradata to Oracle. The AIM project team also developed a portable application that automated the migration of the 13,000 Teradata tables and views.
AIM delivered the project in two major phases:
Phase 1:The initial effort focused on uncovering and gathering all Teradata domain knowledge throughout the organization. AIM conducted more than 20 one-on-one and group interviews and numerous follow-up interviews at multiple organizational levels in this effort. Complex data structures, data analysis, business processes and overall data architectures were discovered and documented in an iterative fashion.
Phase 2: Building on this knowledge, AIM delivered two powerful solutions:
- Confluence WIKI – AIM created a knowledge base and information repository for all processes, tools, the requirements associated with migrating databases from Teradata to Oracle, and more. The team built subsites within Confluence so that employees could post questions, share knowledge and interact more frequently on customer analytics.
- Data Migration Automation Tool – AIM developed an application to automate the migration from Teradata to Oracle. More than 13,000 database objects, constituting 99% of all objects, were migrated leveraging the tool. For larger database objects (i.e. those surpassing 250 gigabytes of storage), ETL workflows were developed, tested and implemented to migrate the data.
To minimize disruption during the migration, most existing business processes were left unchanged. One exception to this is related to security. Data security and access management between Teradata and Oracle necessitated the building of new workspaces for the sharing of data and developing of new processes for accessing data.
Confluence is now the repository for the tribal knowledge formerly stored in employee’s memories and files, with employees utilizing the wiki for everything they need to know regarding which tools to use, how to query the customer analytics environment, and much more. When the Confluence Wiki was released, employees immediately started contributing to it and sharing additional documentation on business processes. This powerful mechanism not only enables employees to ramp up on Exadata quickly, it also ensures business continuity when employees leave the organization.
The data migration automation application was streamlined and developed in three weeks, saving the organization from having to rely on a massive manual undertaking that might have required 20–30 consultants and a year or more to migrate the database objects using an ETL approach. With an application to automate the process, the time invested was trimmed to weeks.
Finally, the data migration strategy project took place during a major transformation at the organization—a revamping of its customer loyalty rewards program. As the organization was busy refining the way it approaches a core aspect of its business, it was relieved not to have to manage the AIM team or the analytics data migration project. Ultimately, AIM’s deep experience with data and analytics enabled the customer to focus its resources in the right areas and avoid disruption during a time of major organizational change.