“Many man-made systems, including ICT systems, have positive feedback loops that cause certain local events to propagate and create extreme global behaviors. The extreme behaviors, especially unplanned downtime, become more common than stakeholders can accept. These outliers are modeled by probability distributions with thick tails. Unfortunately, classical methods for risk analysis based on predictions of future events tend to underestimate or ignore extreme global behaviors in complex adaptive ICT systems, even though these events may very well dominate the overall risk to stakeholders.” RiskDesignArchitectureComplexitySystems ThinkingAntifragileIct Book:Anti-fragile ICT Systems Source: Anti-fragile ICT Systems
“Two examples illustrate the redundancy principle. First, when a virtual machine fails in a cloud-based system, an identical instance is started automatically. Second, a critically important system should have at least one secondary backup system that runs in parallel with the primary system to ensure a safe fallback. Leading up to the next principle, we note that the secondary system should differ from the primary system to avoid both failing for the same reasons.” ComplexitySoftwareSystems ThinkingAntifragileIct Book:Anti-fragile ICT Systems Source: Anti-fragile ICT Systems