It is often normal for scientists to use IT services every day without appreciating the details of their architecture and operational aspects. Symmetrically, often, IT experts, while providing reliable services to the scientific community, do not have the opportunity to dive into the usage patterns of their infrastructures. These two lectures aim to bridge the gap between service users and service providers considering a particular service of the CERN portfolio, SWAN, which providing an interface for web based data analysis, federates several other production CERN services.
The first lecture is dedicated to the general description of the SWAN service, its architecture and components. Concepts, which are generic to any service, will be stressed and distilled in form of general knowledge applicable in areas also not related to SWAN. Special attention is dedicated to the way in which it is packaged in containerized form and how this impacts its operation. Basic concepts of synchronized and shared storage; software distribution, interactive and mass processing are reviewed also through concrete examples collected during the first three years of the service.
The second lecture is oriented towards interactive web based data analysis. The attendees are exposed to the basic principles which lead to a statistical interpretation of a dataset, focusing on examples coming from the collider physics and accelerator physics sectors as well as more traditional big data analysis related to IT infrastructure monitoring. Concepts such as identification of datasets, their cleaning and their effective processing are explored according to the state of the art technology trends.