A Leading Indian Asset Management Company.
The Firm Manages Investments For Millions Of Retail And Institutional Investors Across Equity, Debt And Hybrid Schemes. Its Investor Servicing Platform Processes Purchases, Redemptions And Statements Daily, With Every Transaction Carrying Settlement Deadlines That Cannot Slip.
The Platform Had Been Built Over Fifteen Years As A Single Application With Tightly Coupled Modules And A Database Nobody Fully Understood. Small Changes Required Full Regression Cycles, And Most Engineers Who Wrote The Original Code Had Long Since Left.
Problem Statement
Despite Consistent Business Growth, The Firm Struggled With:
- 01
A Monolithic Codebase
Every Change Required Regression Across Modules That Had No Documented Boundaries.
- 02
Undocumented Data Model
Schema Logic Lived In Stored Procedures Written By Engineers Who Had Departed.
- 03
No Safe Cutover Path
Settlement Deadlines Left No Window Long Enough For A Single Migration Event.
Proposed Solution
The Engagement Delivered Modernisation In Phases Around:
- Mapped The Existing Application And Data Model Before Any Code Or Schema Was Touched.
- Decomposed The Monolith Into Services Along Business Boundaries Rather Than Technical Convenience.
- Modernised The Database With Documented Schemas Replacing Logic Buried In Stored Procedures.
- Ran Old And New Systems In Parallel With Automated Reconciliation Comparing Every Transaction Output.
- Migrated In Waves By Function, Proving Each One Fully Before The Next Began.
- Built Automated Test Suites Covering Transaction Paths That Previously Relied On Manual Regression.
Modernising Without A Single Big Cutover
Each Function Moved Independently, Running Alongside The Legacy System Until Reconciliation Confirmed Identical Results. No Wave Required A Weekend Outage, And Any Wave Could Be Reversed Without Affecting The Ones Already Completed.
Parallel Running With Automated Reconciliation Proved Correctness On Live Data Before Any Function Switched Over.
Service Boundaries Drawn Along Business Functions Allow Teams To Change One Area Without Regressing Another.
Documented Schemas Replaced Stored Procedure Logic That No Current Engineer Could Confidently Explain.
Wave Based Migration Kept Every Step Reversible, So No Single Failure Threatened The Whole Programme.
Result :
A Platform The Team Can Change Again
Investor Transactions Now Run On Modern Services With Documented Interfaces, And Changes Ship Without Full Regression Cycles. The Migration Completed Without A Single Missed Settlement Deadline Or Reconciliation Discrepancy.
Settlement Deadlines Missed
Migration Waves Completed
Faster Change Delivery
Transaction Throughput
Lessons Learned
The Engagement Highlighted Three Lasting Takeaways:
-
Map Before Moving
Undocumented Systems Hide Their Real Boundaries Until Someone Draws Them.
-
Prove With Parallel Running
Reconciliation On Live Data Answered Questions No Test Environment Could.
-
Waves Beat Weekends
A Reversible Migration Removes The Pressure That Causes Big-Bang Failures.
TECHNOLOGIES - TOOLS USED
Modernisation Used Application Mapping And Dependency Analysis Tooling Ahead Of Decomposition Into Containerised Services With Documented APIs. Managed Databases, Reconciliation Engines And Automated Test Suites Ran Alongside The Legacy Platform Throughout, With Deployment Pipelines Delivering Each Wave Independently.
- Application Dependency Mapping
- Containerised Services
- API Gateway
- Managed Databases
- Schema Modernisation
- Data Reconciliation Engine
- Automated Test Suites
- CI/CD Pipelines
- Infrastructure As Code
- Observability Stack
- Migration Orchestration
CONCLUSION
Legacy Systems Are Rarely Replaced Because The Risk Of Cutover Exceeds The Cost Of Standing Still. Migrating In Reversible Waves With Live Reconciliation Removed That Risk, And The Firm Reached A Platform Its Own Engineers Can Change Confidently Without Ever Pausing Investor Servicing.
