Cloud Acceleration Pyramid
The Cloud Acceleration Pyramid outlines the four different types of strategies application developers and their teams should adopt to ship cloud-native software faster. The pyramid also shows the type of organizational focus required for each type of strategy.
Prevention strategies such as unit testing are frequently started as grass-roots efforts by developers. At the top of the pyramid, investing in response readiness requires organization-wide discipline. In between, strategies such as observability and proactive mitigation are a mix of organic adoption with some strategic support.
1. Response Readiness
Ensure that engineers are prepared to fix an issue, any time.
2. Proactive Mitigation
Limit the impact of an issue to as few end users as possible.
3. Observability
Expose the internal state of a system through telemetry, enabling engineers to detect and troubleshoot issues.
4. Prevention
Identify and prevent issues from entering production.
1. Response Readiness
Cloud-native systems live or die by readiness to respond. Engineers need to be armed with the right data, information and tools to fix an issue at any time, especially as the speed of change and levels of complexity accelerate. Enabling response readiness relies on feedback loops, proactive preparation based on lessons learned, and a culture of blameless postmortems to prevent repeat events.
2. Proactive Mitigation
Proactive mitigation is designed to minimize risk when adopting complex, cloud-native microservices architectures. Proactive mitigation strategies and progressive delivery techniques, such as canary releases and blue/green deployments, let you roll out changes incrementally, limiting the scope of any failures to a minimum, while avoiding negative system-wide effects and continuing to ship at speed.
3. Observability
Observability offers a window into operational service level indicators and business KPIs by acting as a system-health monitor through telemetry. In cloud-native architectures, observability has grown more challenging as distributed, modular systems are harder to observe. Adopting an observability stack that provides service-level metrics, monitoring, logging, and distributed tracing, engineers get the data they need to detect and troubleshoot issues.
4. Prevention
Prevention strategies are employed to identify problems and issues and catch bugs early — before they enter production and when they are easiest and least expensive to address. In the cloud-native software development space, prevention is a common cloud acceleration strategy built around creating high-fidelity, realistic and reproducible development and test environments.