Innovations in Serverless & Event-Driven Solutions • Ben Ellerby & Julian Wood • GOTO 2024

Modern cloud transformation requires a gradual, iterative approach rather than a “big bang” migration - using techniques like the strangler fig pattern to slowly decompose monoliths

Platform teams should focus on enabling streamlined feature teams by providing reusable building blocks, constructs, and guardrails while avoiding becoming bottlenecks

Event-driven architectures are seeing a resurgence, but require careful consideration of event schemas, discoverability, and data modeling upfront to be successful

Organizations need clear business-aligned north star metrics (like time-to-value, MTTR, deployment frequency) to measure modernization progress, not just technical goals

Conway’s Law and Team Topologies principles should guide organizational structure - align teams around business domains rather than technical capabilities

Psychological safety and learning opportunities are critical for teams adopting new technologies and ways of working - can’t just focus on technical changes

Cost optimization (FinOps) should be a shared responsibility between platform and feature teams, with platform providing tooling and insights

Temporary data duplication between services can be acceptable if it reduces coupling and complexity, contrary to traditional wisdom

Breaking down modernization into vertical slices helps maintain productivity and show incremental value rather than long-running programs

Edge computing and globally distributed applications are emerging trends pushing more compute and data capabilities toward the edge