Cloud Adoption for Mission-Critical Workloads

Chapter 3. Always On Guiding Principles

Everything breaks.

This statement is a reminder that all systems fail and that it is essential to plan for all types of failures to minimize their impact and ensure continuity of service. The concept that everything will inevitably fail and the importance of planning for that failure is a common principle in many fields, not only IT. The idea is that by anticipating and planning for potential problems, we can design and build more reliable systems.

This principle is also known as designing for failure (DFF) which involves designing and building Always On services that can detect, tolerate, and sustain failures efficiently while transparently isolating the affected components without any human intervention. DFF serves as a guiding principle for decision making and execution, helping to ensure uniformity in methods and practices. This allows architects designing for Always On to apply this principle across every layer of the technology stack and operational model. It also enables a seamless and integrated approach to design, development, and deployment throughout all the components and elements that make up a mission-critical business service.

In the following section, I provide an overview of the cloud deployment methods that are compatible with an Always On approach with architectural principles that must be considered when designing and deploying mission-critical applications in the cloud, ensuring that they can make informed decisions ...