Strangler Fig vs. Big Bang: Choosing the Right Legacy Modernisation Strategy
Rewriting legacy systems is high-risk. The strategy you choose, incremental migration or full rewrite, determines whether modernisation accelerates progress or stalls it.
Modernisation Is a Risk Management Problem
Rewriting a legacy system is one of the highest-risk decisions in software engineering. It is not just a technical challenge. It is a business-critical transition that affects delivery timelines, operational stability, and customer experience.
The strategy you choose determines whether progress accelerates, or stalls.
Why “Start from Scratch” Often Fails
The idea of rebuilding a system from the ground up is appealing. The existing codebase is difficult to maintain, documentation is incomplete, and a modern stack promises cleaner architecture. But most full rewrites underestimate one thing: the depth of the existing system.
Legacy platforms contain years of accumulated business logic, edge cases, and implicit decisions that are rarely documented. Rebuilding that knowledge from memory is far harder than expected.
The result is a familiar pattern. The new system takes longer than planned, delivers less functionality than required, and the old system continues to run longer than intended.
The Strangler Fig Approach
An alternative is incremental replacement. The strangler fig pattern gradually replaces parts of the legacy system while keeping it operational.
Instead of rebuilding everything at once, teams identify well-defined pieces of functionality, often at system boundaries, and reimplement them as independent services.
Traffic for those functions is then routed to the new implementation. Over time, the legacy system handles less and less responsibility until it can be fully retired. This approach reduces risk because there is no single point of failure. At every stage, the system remains usable.
Managing the Boundary Between Old and New
During incremental migration, the old and new systems must coexist. This introduces a critical design concern: how to prevent legacy complexity from leaking into the new architecture.
An anti-corruption layer provides that boundary. It translates between legacy data models and new domain models, allowing the modern system to evolve independently without inheriting outdated assumptions. Without this separation, the new system risks becoming a replica of the old one.
The Data Migration Challenge
In most modernisation efforts, the database is the hardest component to evolve. Application code can be rewritten. Data must be preserved.
Different strategies exist depending on system constraints. Some teams adopt dual-write approaches, where changes are written to both legacy and new schemas during transition. Others replicate data into a new structure and gradually shift read traffic. In more complex systems, historical events are replayed to reconstruct state in a new model.
Each approach introduces trade-offs in consistency, complexity, and operational overhead. The right choice depends on how critical the data is and how much downtime the system can tolerate.
When a Full Rewrite Makes Sense
Despite its risks, a full rewrite is not always the wrong choice. Smaller systems with limited scope, well-understood functionality, and strong test coverage can often be rebuilt faster than they can be incrementally migrated.
Short-lived platforms or products with minimal dependencies are also better candidates for a clean rebuild. The key is predictability. If the system can be fully understood, scoped accurately, and delivered within a controlled timeframe, a full rewrite may be justified.
Choosing the Right Strategy
There is no universal answer. Incremental migration prioritises stability and continuity. Full rewrites prioritise simplicity and long-term clarity. The decision depends on factors such as system size, complexity, business criticality, and tolerance for risk.
In most large-scale systems, incremental approaches provide a safer path.
Final Thought
Legacy systems are not just technical artefacts. They are repositories of business knowledge. Modernisation is not about replacing them as quickly as possible. It is about evolving them without losing what they already do well.
Planning a Legacy Modernisation?
Intagleo Systems helps organizations design modernisation strategies, migrate critical systems safely, and build architectures that support long-term evolution.