Cloud service providers offer thousands of managed services promising to reduce operational burden and accelerate development. But which services actually deliver value, and which ones create expensive lock-in? As engineering teams face pressure to move fast while maintaining flexibility, understanding these trade-offs becomes critical.
This session examines a practical framework for evaluating AWS services based on real-world engineering constraints. The discussion organizes services into four categories along a spectrum from Simple Primitive tools, such as EC2 and S3 to high-lock in services like Bedrock and SageMaker which use proprietary abstractions that are time consuming to escape.
The framework emphasizes core principles: the “airplane test” for local development, blast radius considerations when services fail, and the importance of exit strategies before adoption. If a managed service genuinely reduces operational burden without creating vendor dependence, it may be worth the trade-off. But most services require careful evaluation against these criteria.
The session will demonstrate how to apply this spectrum when making infrastructure decisions, showing when to stay with simple primitives on the left and when to selectively adopt abstractions further right. For teams prioritizing flexibility and local development velocity, the framework provides a practical decision-making tool for navigating cloud provider offerings.