Storage and databases for big data
Tomas Skripcak, Uwe Just, Ida Schönfeld, Esther G.C. Troost, Mechthild Krause · 2019
One of the main objectives for translational research in radiation oncology throughout the world is to accelerate the application of health care innovations into day-to-day care of the cancer patient. Despite the fact that big data in radiation oncology is driven by very different sources, it is very likely that the industrially established big data technologies can present key solutions to fulfil the needs of translational research as well. The traditional, hypothesis-driven approach for generation and validation of medical evidence for clinical practice is the conduction of randomized clinical studies. In hypothesis-driven clinical research, a hypothesis is formulated at the very beginning of data workflow. The big data environment should be able to satisfy diverse needs for data modeling and storage of different data-driven analytical projects. In big data environments, distributed file systems are designed to store data collections reliably in a cluster consisting of a large number of nodes, typically operating on standard commodity servers with directly attached storages.