A comprehensive survey on internet outages

Internet outages are inevitable, frequent, opaque, and expensive. To make things worse, they are poorly understood, while a deep understanding of them is essential for strengthening the role of the Internet as the world's communication substrate. The importance of research on Internet outages is demonstrated by the large body of literature focusing on this topic. Unfortunately, we have found this literature rather scattered, since many different and equally important aspects can be investigated, and researchers typically focused only on a subset of them. And, to the best of out knowledge, no paper in literature provides an extensive view on this important research topic. To fill this gap, we analyze all the relevant facets of this important research topic, stepping from the critical review of the available literature. Our work sheds light on several obscure aspects such as, for example, the different challenges considered in the literature, the techniques, tools, and methodologies used, the contributions provided towards different goals (e.g., outage analysis and detection, impact evaluation, risk assessment, countermeasures, etc.), the issues that are still open, etc. Moreover, it provides several innovative contributions achieved analyzing the wide and scattered literature on Internet outages (e.g., characterization of the main causes of outages, general approach for implementing outages detection systems, systematic classification of definitions and metrics for network resilience, etc.). We believe that this work represents an important and missing starting point for academy and industry to understand and contribute to this wide and articulate research area.