What Comes After Microservices? • Matt Ranney • YOW! 2016

Microservices architecture at scale (2000+ services) introduces significant challenges around composability, coordination and cross-cutting changes

Developer convenience initially drives microservices adoption, but becomes increasingly difficult to maintain as systems scale up

Main challenges with large-scale microservices:

• Cross-service transactions are complex
• Storage and schema changes across services
• Debugging and profiling distributed systems
• Coordinating releases across teams
• Internal self-inflicted DDoS issues

RPCs between services create tight coupling and don’t compose well for complex workflows

Proposed solutions:

• Event-driven architectures over RPC
• Saga pattern for distributed transactions
• Storage scopes for testing schema changes
• Composable event processors with orchestration layer
• Vendor-neutral infrastructure for event processing

Organization scaling benefits of microservices remain valuable despite technical challenges (Uber grew from 200 to 2500+ engineers)

Need better tools and abstractions for:

• Local development with subset of services
• Safe production data access
• Cross-service schema changes
• Service composition and orchestration

Future direction points toward event-driven architectures with better composability rather than direct service-to-service RPCs

Balance needed between complete decoupling (microservices) and monolithic coupling - neither extreme is optimal at scale