Database Migration from Oracle to PostgreSQL on AWS platform

Empowered a large pharmaceutical company to improve the TCO, scalability and query performance by building a data lake on AWS and migrating Oracle DB applications to PostgreSQL Database

Augmented-Analytics-banner

About the Client

The client is a multinational pharmaceutical company based in India. The client deals in manufacturing and selling active pharmaceutical ingredients and pharmaceutical formulations in India and the US. Considered as one of the most reputed brands in India, the client has expanded with joint ventures and acquisitions in the last two decades.

The Business Challenge

The client wanted to migrate their existing Oracle-based (EDW, GDW) and Tableau reports to open-source software and tools. With their new business model, the additional requirement of dashboards, reports, and enhancements around existing dashboards, data volume increased greatly.  With online access provided to every sales representative, the user base rose to 9000+.  Therefore, instead of investing further into on-premise infra with licenses, the client wanted to move to On-Cloud open source platform expecting to keep TCO in control with rising volume and requirement. Scalability, performance, cost, and improvement of batch job processing were the key factor for the analytics modernization.

What Aptus Data Labs Did

Our team analyzed the client’s existing infrastructure, platform, tools, and taken-up tasks and:

  • Designed and built a data lake architecture covering AWS infrastructure, AWS services, and open sources software based on the pros and cons of platforms along with cost implications
  • Prepared data & application migration strategy and data pipeline for automation
  • Setup and configured enterprise querying and open source-based reporting platform
  • Completed design, development, migration, and testing of the migration
  • Provided 24x7 support

The Business Impact Aptus Data Labs Made

Our approach helped the client to

  • Reduce cost of system stack by using “Open source yet industrial grade” platforms
  • Reduce time required for analysis, analytics, and reporting on “Structured and Unstructured data”
  • Improve and accelerate analytics-driven decisions
  • Prepare ground and infra for AI/ML and Advanced Analytics

The Business and Technology Approach

We followed a phased approach for the migration:

  • Assessment: We conducted a thorough assessment of the application, compatibility, and architectural assessment
  • Architecture & Services – we built the architecture, framework and selected the AWS services and open source tools
  • Schema Migration: We used tools, such as Ora2pg, DMS, SCT, and in-house developed scripts for migration, which enabled faster schema migration. Since there were challenges to migrate some of the schemas like procedures, functions, and scripts were manually converted.
  • Functional Testing: Before proceeding further, it was important to test the converted schema on a sample dataset. We loaded a sample dataset in both Oracle and PostgreSQL databases test environment and tested to ensure the test results are identical. Any issues encountered during this functional test were addressed immediately.
  • Performance Testing: The application might see some difference as few of the Oracle built-in transactions or features functionality would be slightly different in PostgreSQL. The application was tested and any differences revealed were addressed and fixed.
  • Data Migration: Data was migrated in two phases. One using a snapshot for a cut-off period and then in the second phase did incremental loads.

Architecture

Tools used

  • AWS Platform
  • Ora2Pg
  • Oracle
  • PostgreSQL
  • AWS Schema Conversion Tool (AWS SCT)
  • AWS Database Migration Service (AWS DMS)
  • Our data accelerators

The Outcome

With spiraling license fees from Oracle, the potential upside in terms of cost savings was easy to appreciate. This engagement not only provided the client with the evidence of skills and expertise that Aptus Data Labs could provide, it also made it an obvious decision for the client to award Aptus Data Labs the contract for 24/7 Production Support.

Some of the benefits to the client were:

  • Saving the cost of 9000+ oracle end users
  • Reducing the cost of support and streamlined 24x7 support through alert and automation
  • Usage of hot, warm, and cold data storage concept for fast processing within a defined window and effective backup & recovery
  • Migration of 900+ tables, 600+ stored procedures, 3.2TB data and 200+ daily job scripts

Related Case Studies

Download Case study

Explore more of our client stories

Download Case study