Strangler Fig vs Big Bang Migration

When replacing a legacy system, the two main strategies are the Strangler Fig pattern (incremental migration) and Big Bang migration (complete rewrite and cutover). The choice significantly impacts risk, timeline, and team productivity.

Strangler Fig Pattern

Named after strangler fig trees that gradually envelop their host, this pattern incrementally replaces legacy functionality. A routing layer (often an API gateway) directs traffic: new features go to the new system, legacy features continue on the old system. Over time, more routes shift to the new system until the old system can be decommissioned.

Big Bang Migration

The team rewrites the entire system from scratch, then cuts over from old to new in a single deployment event. The old system is turned off, and the new system takes over.

Why Big Bang Rewrites Fail

History is littered with failed big bang rewrites. The problems are predictable: the rewrite takes longer than expected, the old system continues evolving (creating a moving target), scope creep adds features to the rewrite, and by the time it is "ready," the business has changed.

The Strangler Fig avoids these traps by delivering value incrementally and keeping the old system as a safety net.

The Routing Layer

The key infrastructure for Strangler Fig is a routing layer that can direct requests to either the old or new system based on URL path, feature flag, or header. This is typically an API gateway, reverse proxy, or load balancer.

For more migration patterns, see our system design interview guide and architecture concepts. Explore pricing for practice.

The Bottom Line

Default to the Strangler Fig pattern for any production system migration. Reserve big bang for small, non-critical systems where the rewrite can be completed quickly and the risk of failure is acceptable.