Strangler pattern
Generated by GPT-5-miniExpansion Funnel Raw 89 → Dedup 0 → NER 0 → Enqueued 0
| Strangler pattern | |
|---|---|
| Name | Strangler pattern |
| Type | Software design pattern |
| Introduced | 2000s |
| Related | Martin Fowler, Enterprise Integration Patterns, Microservices architecture, Service-oriented architecture |
Strangler pattern The Strangler pattern is a software refactoring approach for incrementally replacing or modernizing an existing legacy system by routing functionality to a new implementation alongside the old until the legacy system can be retired. It emphasizes incremental delivery, risk reduction, and continuous operation during transformation projects led by teams at organizations such as Netflix, Amazon (company), Spotify, and Google.
Overview
The pattern emerged from practices popularized by Martin Fowler and draws on ideas from Enterprise Integration Patterns, Branch by Abstraction, and techniques used in large-scale transformations at Microsoft, IBM, and Oracle Corporation. It typically places a façade, proxy, or router in front of a legacy monolithic application to divert selected requests to a new microservice-based implementation while leaving other requests to the original system. Adoption stories cite deployments in contexts like Bank of America, Goldman Sachs, and Capital One where compliance-sensitive workloads required gradual migration.
Context and motivation
Organizations pursue the pattern when refactoring a system developed over decades by teams at General Electric, Siemens, Siemens AG, or Boeing that cannot be replaced in a single cutover. Drivers include scaling constraints faced by companies such as Facebook, Twitter, and LinkedIn; technology debt noted in reports from Gartner and Forrester Research; and regulatory pressures encountered by institutions like Deutsche Bank and HSBC. The technique reduces risk compared with big-bang rewrites seen in projects at NHS and U.S. Department of Defense by enabling progressive verification, feature parity, and rollback capability.
Implementation patterns
Common implementation choices include proxy-based routing using tools employed by NGINX, Envoy (software), or HAProxy; API gateway strategies similar to those used by Kong (company), Apigee, or AWS API Gateway; and database migration tactics inspired by Event Sourcing and CQRS implementations at Uber and eBay. Teams often combine feature toggles pioneered at Facebook with service mesh capabilities developed by Istio or Linkerd. For data migration, techniques like dual-write, change data capture used by Debezium, and strangler-style table-by-table migration have been applied in projects at PayPal and Stripe.
Variants and related approaches
Variants include façade-first approaches used in Oracle Corporation integrations, database-first strangling practiced at JPMorgan Chase, and front-end strangling applied by Airbnb and Pinterest to migrate user interfaces. Related approaches include Anti-Corruption Layer from Domain-driven design, Branch by Abstraction advocated in ThoughtWorks case studies, and Parallel Run strategies used by NASA and European Space Agency for mission-critical systems. Contrasting strategies include big-bang rewrites seen at Boeing and Rolls-Royce and rip-and-replace migrations used by Walmart.
Advantages and disadvantages
Advantages promoted by proponents at ThoughtWorks and Gartner include reduced deployment risk, continuous business value delivery for clients like Salesforce and SAP, and incremental testing with organizations such as Red Hat and Canonical (company). Disadvantages observed in case studies from McKinsey & Company and Accenture include prolonged maintenance of dual systems, coordination overhead across teams at Intel and AMD, and complex data consistency challenges encountered by Citigroup and Morgan Stanley. Security and compliance implications require controls familiar to ISO standards and regulators like SEC or European Central Bank.
Real-world examples
Notable real-world adoptions include migration projects at Netflix where monolith-to-microservice transition inspired public talks; the backend evolution at Amazon (company) that influenced service decomposition practices; and front-end incremental rewrites reported by The Guardian and BBC. Financial institutions such as HSBC and Goldman Sachs have documented table-by-table and service-by-service strangling for trading platforms. Open-source and vendor tools used in these efforts include Kubernetes, Docker, Terraform (software), and orchestration frameworks used by Red Hat and VMware.
Best practices and considerations
Recommended practices endorsed by practitioners at Martin Fowler's community, ThoughtWorks, and enterprise architects at McKinsey & Company include establishing routing and observability with Prometheus and Grafana, enforcing API contracts through OpenAPI Specification and gRPC, managing schema migrations with tools like Flyway and Liquibase, and coordinating releases with CI/CD pipelines using Jenkins, GitLab, or GitHub Actions. Governance and stakeholder alignment involving teams across CTO offices and compliance functions at European Commission-regulated firms help manage risk. Plan for decommissioning costs, runbooks inspired by ITIL practices, and continuous testing strategies used by Google SRE teams.