This article discusses the Architecture Advice Process as a method for decentralizing architectural decision-making, moving away from traditional 'ivory tower' or 'hands-on architect' models. It emphasizes empowering teams to make decisions by seeking advice rather than permission from affected parties and experts, fostering shared understanding and faster feedback loops. This approach addresses the scalability challenges of centralized architecture by distributing responsibility and accountability across the organization.
Read original on InfoQ ArchitectureTraditional architectural practices often concentrate decision-making authority in a few named architects. This model becomes a significant bottleneck as systems and development practices evolve, leading to architects becoming 'flow-blockers'. Architectural decisions happen constantly and everywhere, making it impossible for a small group to possess all necessary context, information, and accountability across an entire organization.
Scaling Bottlenecks in Traditional Architecture
Centralized architectural decision-making often leads to bottlenecks, slow feedback loops, and a lack of accountability for those implementing the architecture. It struggles to keep pace with the dynamic nature of modern software development and distributed systems.
The Architecture Advice Process decentralizes decision-making by empowering any team member to make architectural decisions, provided they seek advice (not permission) from two key groups: those who are affected by the decision and those with relevant expertise. This transfers both responsibility and accountability to the decision-maker, fostering a culture of knowledge sharing and collective ownership.
Consider a team choosing a new platform for a service (e.g., serverless lambdas vs. Kubernetes microservice). A systems architect might advise against lambdas due to perceived startup latency or state management complexities. Under the advice process, the team can still choose lambdas, understanding the architect's concerns but remaining accountable for their choice. This shifts the dynamic from a hierarchical approval process to one of informed decision-making and shared learning.