A Socratic method for validation of measurement-based networking research

Nearly three decades of Internet measurement has resulted in large-scale global infrastructures used by an increasing number of researchers. They have examined various Internet properties in areas such as network infrastructure (routers, links), traffic (measurement at packet, flow, and session level) and applications (DNS, Web, P2P, Online Social Networks, etc.) and presented results in diverse venues. Key related topics like security and privacy have also been explored. There is however a lack of clearly articulated standards that reduce the probability of common mistakes made in studies involving measurements, their analysis and modeling. A community-wide effort is likely to foster fidelity in datasets obtained from measurements and reused in subsequent studies. We present a Socratic approach as steps towards a solution to this problem by enumerating a sequence of questions that can be answered relatively quickly by both measurers and reusers of datasets. To illustrate the applicability and appropriateness of the questions we answer them for a number of past and current measurement studies.