Notions on robustness exist in many facets. They come from different disciplines and reflect different worldviews. Consequently, they contradict each other very often, which makes the term less applicable in a general context. Robustness approaches are often limited to specific problems for which they have been developed. This means, notions and definitions might reveal to be wrong if put into another domain of validity, i.e. context. A definition might be correct in a specific context but need not hold in another. Therefore, in order to be able to speak of robustness we need to specify the domain of validity, i.e. system, property and uncertainty of interest. As proofed by Ho et al. in an optimization context with finite and discrete domains, without prior knowledge about the problem there exists no solution what so ever which is more robust than any other. Similar to the results of the No Free Lunch Theorems of Optimization (NLFTs) we have to exploit the problem structure in order to make a solution more robust. This optimization problem is directly linked to a robustness/fragility tradeoff which has been observed in many contexts, e.g. 'robust, yet fragile' property of HOT (Highly Optimized Tolerance) systems. Another issue is that robustness is tightly bounded to other phenomena like complexity for which themselves exist no clear definition or theoretical framework. Consequently, this review rather tries to find common aspects within many different approaches and phenomena than to build a general theorem for robustness, which anyhow might not exist because complex phenomena often need to be described from a pluralistic view to address as many aspects of a phenomenon as possible. First, many different robustness problems have been reviewed from many different disciplines. Second, different common aspects will be discussed, in particular the relationship of functional and structural properties. This paper argues that robustness phenomena are also a challenge for the 21st century. It is a useful quality of a model or system in terms of the 'maintenance of some desired system characteristics despite fluctuations in the behaviour of its component parts or its environment' (s. [Carlson and Doyle, 2002], p. 2). We define robustness phenomena as solution with balanced tradeoffs and robust design principles and robustness measures as means to balance tradeoffs.